Remote Laboratory Approaches
“Education is not filling a bucket but lighting a fire.”
– William B. YeatsChapter contents
11.2 Typical System Architectures
11.3 Remote Access Technologies
11.4 User Observations
11.5 Converting from Traditional Labs
11.6 General Requirements
11.7 Software Requirements
11.8 Overall Design Considerations
11.9 Jump Starting Remote Labs
The operation of a typical remote labs software package is given in Appendix D.
Student surveys tend to indicate that most learners find a well-constructed remote lab (with a simple interface) useful and equivalent to that of a traditional lab.1 However, there are misgivings about the lack of hands-on experiments with instruments and equipment, as well as some conjecture regarding whether a remote lab can accurately reproduce the feeling of working in a standard lab.2
A study has shown that remote labs encourage students to operate experimental equipment for a longer period of time than in a traditional lab, and learning outcomes appear to be improved.3
Remote instrumentation is becoming widely used with notable examples being a telescope (e.g. MIT Haystack Observatory 37m radio telescope), a single crystal X-ray diffractometer (Youngstown State University) and adiabatic flame temperature (Columbia University).4 To ensure the long term sustainability of the instrumentation, requirements here would include appropriate levels of staffing, marketing of the facility, dissemination of research results, regular design and implementation of new projects and regular updates of the equipment itself. Most remote students found the experiments useful with the same exposure as local students, although the lack of immediate staff support was a disadvantage.
The following section examines typical remote lab system architectures. Next remote access technologies are detailed. User observations on the best way of building a remote lab are described. The suggested conversion process from traditional to online remote labs is then discussed. General and then software requirements are assessed followed by overall design considerations. The chapter is completed with a discussion on the lack of evidence of successful sustainable remote labs followed with a conclusion.
11.2 Typical system architectures
The typical architecture of a remote lab comprises a learning management system (LMS) which learners interfaced to through the web server.5 The LMS then connects through to a series of lab servers which allow access and control of the actual lab equipment. Each lab server is connected to a measurement server which is physically connected to lab equipment (either hard wired or through either Ethernet, RS-232, USB or GPIB communication buses). The main services provided by the remote lab modules include a demonstration of the experiment from instructor to the student, visualization of the experiment by the student, control of an experiment by the studentand, finally, construction of an experiment by the student.
Figure 11.1: Typical remote lab system architecture. There are essentially four main components to a remote lab:
• The learner’s PC with client software.
• A central server (with scheduling, web, database services, web conferencing software).
• A lab server with appropriate software.
• Measurement server, equipment and instrumentation connected to the lab server.
There are new opportunities to learn not only from handling equipment but in interacting together in collaborative ways on remote labs with students located from different locations.6 Students can learn not only from use of the lab tools but in interacting with each other with divergent perspectives and also in shared knowledge building. There are a number of different approaches to collaboration: students can collaborate in face-to-face groups on one location, individually in different locations or in groups in different locations.
Another proposed remote lab architecture used at the University of Toronto referred to as an eLaboratory comprised the following modules: remote experiment, engineering portal (for storage of student’s documents–akin to a lab notebook), telepresence (streaming video of the lab and collaboration with others or the lab assistant), application publishing (allowing access through remote desktop protocol to other programs) and scheduling (for the students to access the labs without conflict).7Typical system architecture of a remote lab
• The client application is a Java-based web browser which can access the internet.
• The resource manager is based around a web server to respond to HTTP requests from clients and for authenticating all users. In addition, it has a database server for storing all data relating to users, equipment descriptions and results of the lab sessions. The final component is a scheduling server for experiments and equipment and for allocating experiments.
• The third component comprises the computer interface and the actual instruments and actuators.
The associated software proposed was as follows:
• The graphical user interface (or GUI) resides on the client, allowing the user to the interface to the lab. This should be open to a variety of web browsers, thus achieving platform independence.
• The second layer is the web application which processes requests from the GUI and responds with the requested data. A variety of different technologies can be used here, ranging from Common Gateway Interface (CGI), PHP, JavaServer Pages (JSP), ASP.NET and ASP.
• The third layer is the lab agent which responds to requests from the web application and extracts the necessary information from the instrument or writes information to an actuator. At the end of the experiment, the experimental data is stored in a database.
• The fourth layer comprises the instruments, actuators and cameras software driver interfaces which interact with the lab agent.Typical components of a remote lab
Another suggested approach for a remote lab comprised:10
• Web and Database server containing user accounts, experimental documentation and monitor of students’ experimental activities.
• Collaborative Server allowing for a synchronous web conferencing experience.
• Experimental computer connected to the switching matrix and equipment.
• Switching matrix allowing for remote switching of the various experimental components.
• Experimental equipment which is connected to the experimental computer and switching matrix.A suggested architecture for remote labs
There are many architectures that have been built up for remote engineering labs, mostly based around a client (at the student’s end) and a lab server connected to the experiment and a web server handling the administrative issues.11
A remote lab structured in a three layer model can be visualised with layer 0 being the actual experimentation, layer 1 including the basic modules such as database, data archiving, video server, collaboration, authentication and scheduling. Layer 2 has the visualisation and instructor administration modules. A literature review suggests that the following model could be built up for a typical remote lab:Layer 2
This interface allows for such elements as log in, modifying passwords, scheduling arrangements, retrieving data from previous experiments, initiating the experiment, collaborating with other students and reading information posted by the instructor.Instructor administration tools
This allows the instructor to remotely configure an experiment, post general information for the lab participants and start or stop an experiment.Layer 1
This is to ensure only registered lab participants are allowed to access a specific lab and is based on a database being kept of authorised participants, with their allowed time slots, as well as advising when an experimental slot's time is about to expire.Scheduling
This allows for scheduling of all experiments so that participants can avoid conflict and congestion in booking labs (particularly before the deadline for submission of a lab report!).Video serving
A number of cameras are provided here for access by participants with repositioning of the camera allowed as well as changing the rate of frames per second for the video.Collaboration server
This allows participants to work with each other on a specific experiment using combinations of text chat, audio, video and whiteboard.Data archiving
All data from experiments can be archived in the system database and can be retrieved by the participants.Database
The database module would contain such information as participant's profiles, instructors, list of experiments, scheduling data and archived data.Experimentation
The suggestion is that the ubiquitous LabVIEW web-publishing software can be used for experiments but any of the following technologies could be used for the lab hardware: PCI/USB/Ethernet and GBIB.Concurrency
This allows for multiple simultaneous participants in an experiment with only one user at a time controlling the experiment and the remainder watching.
11.3 Remote access technologies
The history of remote access
There has been a steady movement from Virtual Private Networks (VPNs), which use standard encryption protocols such as internet Protocol Security to create a private tunnel, to the Secure Sockets Layer (SSL) VPN which eliminates the requirement for the installation of specialized VPN client software on the client’s computer.12 Recently, Desktop Virtualization allows for users to access their virtual desktops from a centralized server. The final (major) trend has been the growing use of downloadable mobile apps which allow for secure mobile / remote access to applications and data without the need to open up a web browser session.
Besides Microsoft’s Remote Desktop, there are a number of remote access technologies that could be used for remote labs.13These Mac and Windows compatable packages include:Teamviewer. An excellent package for remote control, this is normally installed on the server side with a client being able to access it, transfer files and chat. There is no direct installation of software on the client side. It includes full encryption based on the https/SSL standard. Clients have complete control of the mouse and keyboard at the same time, thus making it unsuitable for multi-user use. It costs $700 for lifetime commercial use (2009). Response delay is of medium level.
GoToMyPC. he connection between client and server is through an additional server connection, thus making it somewhat slower than other offerings. There is also a need to install software on both client and server. It uses the Advanced Encryption Standard and allows for file transfer, chat, remote sound, file printing and drawing abilities. There is no multiuser ability and software installation on the client side can be problematic. It costs $180 pa (2009). Response delay is high.
LogMeIn. Besides software requiring to be installed on the server (but not the client), this uses high level SSL end-to-end 128-to 256-bit encryption. This provides chat, remote control, remote sound and file transfer. There is a monthly charge of $20pm for LogMeIn Pro. Response delay is high.
11.4 User observations
In order to build a good design for a remote lab, it is important to examine user observations on what is required to build a remote lab.Errors in remote labs
With the different ways of applying instruments (and control equipment) in remote labs, there is the possibility of introducing errors into the measurements as compared to that of the traditional classical labs.14 What makes the issue more challenging is that often there are multiple users accessing the remote experiments simultaneously and the remote lab set up has to be able to handle this situation. A standard electronics lab comprising function generator, digital multimeter, oscilloscope, power supply and switching matrix was analyzed with minimal errors detected in frequency, amplitude, voltages, resistances and currents.System response time
The system response time for various numbers of concurrent users was investigated with the electronic test circuit (including the switching matrix). Responses ranged from 800msecs for 10 simultaneous users with a simple circuit (one component in the switching circuit) to almost 7 seconds with 60 simultaneous users with a complex circuit (3 components connected to the switching circuit). It was suggested that the absolute maximum time a user would be prepared to wait for results would be 7 seconds (but this would be considered an absolute extreme).
A comparison was made between three different web-based LabVIEW control systems based around flow control of a three-tank system (although the live video broadcasting was the same for all three).15 The web-based LabVIEW to web browser control through CGI method delivered the best performance with the shortest data acquisition time (0.4 sec period), whilst the LabVIEW client to LabVIEW server using different datasocket channels (4 seconds period) and web-based LabVIEW to web browser control through ActiveX + Datasocket were the slowest. The only major challenge for the CGI approach is that the data are only downloaded at the end of the experiment and is not visible in real time.Time scheduling issues with remote labs
As opposed to batch-mode experiments, interactive ones require instantaneous control of equipment and real-time measurements but it is difficult to provide open-ended unlimited access with these, as the dem and from a number of users can conflict and each experiment’s time is normally fairly short.16 This means that a scheduling scheme is necessary and this can vary from time slot mode, queuing mode or a hybrid (slotted-queuing) mode. Time slot refers to the duration of the experiment; the user selects her preferred time from a menu indicating open slots and booked slots. While this has the advantage of dedicated access at the user’s preferred time and more flexibility in varying testing conditions for multiple measurements, the disadvantage is inefficient utilization of resources.
The queued method allows for several users to use the equipment with some unpredictability in a short period of time–possibly even concurrently. The advantages here are optimized efficiency (available all the time), increased convenience and the fact that the method is suited to a comparatively large number of experiments with fewer measurements (and measurement conditions changing). The disadvantages are lack of full control, less predictability in exact time of use and an unpredictable and extensive waiting time. A suggested compromise approach (with better efficiency than a time slot) is slotted queuing with a limited maximum number of users per time slot.Importance of reality
Obviously, remote labs being detached from the student working on them have an impact on the student’s feeling of presence.17 The student’s feeling of presence, “lies in the perception of reality”–or the lack thereof. This sense of presence is a good predictor of performance in the lab and of success in achieving the learning outcomes of the experiment. A recommendation here is to ensure the graphical user interface is as realistic and easy to work with as possible. The graphical user interface should not interfere with the student’s perceived contact with the real lab experiment located at the remote location. A further suggestion to improve the experience is to allow the student to work with a proximal lab first, thus helping to establish the feeling of reality when transitioning to work on the remote lab.
If the lab equipment is used more efficiently with many simultaneous users (and a resultant minimal of use of cameras showing real equipment), the student will suffer a reduction in her perception of reality.
As would be intuitively expected, the problem with batch mode of operation is that in a remote lab setting, the feeling of presence is lacking.18
A transparent graphical user interface (GUI) is vital to the execution of a good experiment and to assisting in providing the student with a sense of reality.19Real world interaction to virtual worlds
Interactions in these instances range from totally real world interaction to mixed reality (real and virtual combined) to a purely virtual world.20 Mixed reality could be a proposed new building superimposed on a real world street (often referred to as augmented reality) viewed through a head mounted display. In terms of building in remote labs or other equipment, the Open Wonderl and virtual world is probably optimum because it is open source and is easier to expand.Requirements for integrating labs into professional practice
In working with remote labs, it has been suggested that there are two types of reality.21 The first is initiation or establishment reality, which is the minimum required to convey to the student the impression that she is working with real but remote equipment. The second is maintenance reality, which is lower and provides the stimulus to maintain the student’s belief that she is still working with real equipment. One of the most significant contributors to reality is a live video feed of the lab setup.
Elements that can contribute to the very important professional and industrial (as opposed to academic) context of the remote lab experience include the type of professional setting that places the experimental equipment squarely within the context of industry (e.g. a Programmable Logic Controller working to control a set of motor drives in a plant), real world complexity and ambiguity (e.g. time and cost limitations and collaborative team challenges in working on the experiment), the locus of control devolving to the student as to the conduct and flexibility of the experiment and finally connection to real problems in industry.Good lighting is essential
The University of Deusto undertook a survey of students undertaking the lab on PLDs (programmable logic devices) over the past six years, with 247 completing the survey.22 Overall, the students considered that the remote lab was a useful learning tool. The major irritation was that the webcam was inadequate, mostly due to a lack of light; once this problem was rectified with an improved lighting system, subsequent surveys indicated satisfaction. Remarkably, there were over 1,000 accesses of the remote lab for the 23 students surveyed. 75% of the remote lab accesses were done from outside the university campus with the majority done in the afternoon and night.Student feedback from large-scale trial or remote labs
A survey was conducted of lab sharing between August 2009 and September 2010 with six Australian universities: CQ University, Griffith University, Monash University, University of Adelaide, University of Tasmania and University of Technology, Sydney.23 The remote labs included loaded beams, coupled tanks, FPGA-based digital systems, an iRobot, shake tables, PLCs and the Coldfire real-time operating system. Over 1,000 students undertook the lab sessions, mostly formally but there were also some informal participants. From these, 157 responses to a survey were analyzed for user observations.
An immediate observation was that the highest usage of the labs took place in the days immediately before a lab submission had to be made.
In reflecting on the survey results, the use of remote labs, as compared to that of hands-on physical labs, exhibit a number of practical differences to the students, which included:
• Minimal sensory interaction (such as tactile).
• Individual work vs. group work.
• Flexible time vs. fixed time.
• Stronger reliance on notes against extensive dependence on lab tutors.
The survey responses were illuminating:
• 74% of students had never heard of remote labs.
• Most accessed the labs from home and from university computer labs.
• There was a strong preference for afternoon and evening access to the labs.
• The convenience and flexibility of access to remote labs were appreciated but only where hands-on labs are unavailable.
• Video feedback of the scene was useful in instilling trust into the data sourced from the remote lab.
• The advantages were defined as flexibility (time and location), ability to repeat experiments and ease of use.
• The disadvantages were listed as lack of engagement, technical problems (control and video), less learned than in proximal labs and delay in accessing lab.
• Most wanted more remote labs with only a small percentage (9%) preferring hands-on proximal labs.
A few recommendations suggested for future remote labs were:
• Instructors need to be taught the finer points of running remote labs to make them successful. There are some significant challenges in sharing a few items of equipment with large numbers of students.
• Remote labs should be made compulsory or count towards the final mark and represent a tangible reward.
• Students will learn to accept and work with remote labs as a normal part of their curriculum if they use them more frequently.
• The learning outcomes between remote labs and hands-on labs are different and need to be carefully identified and used.
It is always important in the client server application to keep all time-sensitive real time algorithms located in the server rather than the client due to possible introduction of significant dead times.24
Finally, there was a plea for everyone to join the remote lab community and build it into a very successful resource.Survey of stuydents at Florida Atlantic University
A survey to a limited number of students indicated that even those undertaking online courses preferred the conventional physical lab in which to conduct experiments.25 However, most preferred remote labs to simulations.
Initial research revealed that students valued support in their remote lab through synchronous communications between a remote tutor and students during the lab session.26 This could be extended to asynchronous communications if the labs and support are conducted on an international basis. Application sharing was of critical importance for the tutor to assist the students in their tasks.Constraints in adoption of remote labs
There is often no overwhelming reason to change to remote labs from existing conventional labs.27 Although the number of real labs undertaken may be fewer when people engage with remote labs, for example with a reduction of 30% in the number of labs, this doesn’t provide enough of a reason to change to remote labs. Added to this, reluctance to change is the high level of inertia.
There is a lack of universal standards for remote labs apart from use of LabVIEW software.
There are some unique restrictions such as bandwidth, the need for computer-related skills and concerns about the artificiality of the environment and results (e.g. lower levels of noise).
There are concerns that remote labs are “not real” and thus cannot provide the same level of experience as a conventional lab with overwhelming auditory, visual and kinesthetic senses (and even olfactory, with a burnt out resistor).
Academics who would need to support these labs are not generally motivated by money or organizational requirements and would generally not be overly enthused with a set of remote labs thrust on them but would be personally motivated by contributing to the design and operation of a lab.28 It was thus suggested that an approach may be to construct remote labs which are highly configurable so they could then design the way in which the remote lab would be utilized for their students.Multiple video streams can be demanding
Although this has been discussed to some extent in an earlier chapter on synchronous webconferencing, multiple video streams for a lab can be challenging. Multiple video streams can be demanding on the client, especially in terms of peak CPU utilizations making the client computer unresponsive for heavy loads.29 A suggested approach is to adjust the encoding parameters (frame rate and size) in an iterative way to reduce loading to an acceptable range.
Each video stream process is setup with an encoding profile (video device name, network protocol, port number, encoding mode (constant bit rate or variable bit rate), encoder type (e.g.H.263), frame size, frame rate, constant bit rate value and variable bit rate quality level (0 to 100). In variable bit rate, for a given quality level, the bit rate will vary depending on the complexity of the data stream (e.g. a high rate of change in a scene will result in a higher bit rate). Variable bit rates provide constant quality across all video streams, but bandwidth requirements are unpredictable, whereas with constant bit rates, the video quality will vary but the bandwidth is predictable.
A typical network protocol used is the Real-Time Transport Protocol (RTP). The system loading (e.g. CPU/Memory consumption/bandwidth) is constantly measured and once it reaches a defined threshold, either the frame size (640x480, 320x240, 160x120) or the frame rate (30, 25, 20, 15) is adjusted downwards, until the loading is back in the correct range. If overload is still detected after reducing the frame rate for the first stream, the next video stream will be reduced. Once all streams have been reduced to their defined quality levels, the first stream’s frame size will then be reduced and so on. In an alternative scenario, the frame size can be set as the first parameter to adjust and the frame rate as the second. This algorithm has proved effective in reducing CPU load.Strategies to improve collaboration by rejigging remote lab architecture
Close collaboration between (often, groups of) students and instructors working on a lab experiment can be of enormous help in achieving strong learning outcomes.30 Indeed, collaboration can simply be a serendipitous encounter between students working on different lab experiments but who assist each other.
An initial strategy at University of Technology Sydney (UTS) was to run VMWare on the lab server, thus allowing multiple virtual PCs to operate on one item of hardware. Each student had a unique IP address for each virtual PC. The virtual machine interface is displayed on the student’s remote computer, thus there was no need to install licensed copies of the software on each student’s remote computer. However, whilst multiple students can observe the actual lab equipment through a video feed through each virtual machine, only one can view and operate the control software application. A better approach where control is required on multiple machines is to use the VNC (Virtual Networking Computing) software. VNC, a remote desktop sharing utility, allows for the control software application to be displayed on multiple remote machines.
A reported 40% of students at UTS preferred to use instant messaging (text chat) as the method of communications to their lab tutor. Open source PhpFreeChat was used to create virtual chat rooms for each lab. Overall this has worked well, but an extremely slow response was encountered when interacting with the control application on the server virtual machine.Laboratory management systems
A laboratory management system (LMS) is a useful addition to manage learners, for managing and allocating resources (registration, instructor availability, training resources and delivery of online learning) and a scheduling system for the lab hardware.31 In addition, authoring tools to create lab tutorials, quiz and data analyses are useful features.
The management of an online lab system requires at least the following elements:
• Scheduling of experiments.
• Delivery of assignments.
• Undertaking the remote experiment.
• Publishing of the lab report.
• Interactive discussions between students and instructor.
• Evaluation and assessments.
The preparation of a comprehensive lab report is a key part of every engineering and science student’s education.32 It is always tempting for the hurried instructor to try and short circuit this process by getting students to complete an online quiz to assess their work done in a particular lab; however, it is vital for the student to go through the complete process of writing a comprehensive and competent lab report. This forces them to think logically, analyze the data, and to communicate effectively in relating practice to theory.
Instructors should ensure that students know how their reports will be graded with examples of good reports provided. It is possible if the students achieve a sufficiently high standard in their first three or four lab reports for the instructor to substitute an online quiz for part of the work at this point. It is essential to build in some checks to the assessment process as to whether the student actually undertook the experiment and understood what they were doing. As lab experiments (and writing up the report) can be time consuming, instructors should immediately chase up students who are falling behind with this activity so that they are done in a timely fashion. Trying to do all the labs in one hit at the end of the course makes for a poor learning experience.Observations on great remote labs
It was suggested that the following characteristics make for a truly great remote lab:
• Initially providing theoretical background to the experiment, then a simulation and then finally the remote lab.
• Working hard to achieve a strong feeling of reality and proximity to the remote equipment.
• Allowing for flexible study and more effective time management as they can be done at any time or place.
• Being economically more effective than traditional labs; however instructor’s time and efforts can be the same.33
• Prompt support to students with their remote lab endeavors encourages students to greater efforts as well as increased interest.34Survey at Netlab
A survey was conducted for the well-known remote lab called NetLab at the University of South Australia for two undergraduate courses on Electrical Circuit Theory (second year) and Signals and Systems (third year).35 There was some unease amongst the students regarding working alone with a student from another country so the team sizes were increased to four (two from Australia and two from Singapore). It was found that the team leader role was critical to the success and prior experience with remote labs makes for more enjoyment and satisfaction. There is minimal supervision (as compared to the more supervisors in the classical labs) hence training on the use of the remote labs and a detailed guide for each experiment is absolutely critical. Initial training can be built within the standard lectures, using pre-recorded online videos and in introducing students to remote labs in their first standard lab session.
One issue, which is vitally important, is to ensure absolute reliability of the remote labs at all times, with some sort of backup server which can be brought into play very quickly. Nothing can create more irritation than a remote lab that doesn’t perform as required– especially when the issue is exacerbated by thin technical support.No distractions
A review of a number of remote experiments for remote networks at the Universidad de Valparaíso, Chile suggested the efficiency of remote labs is considerably better than the face-to-face one due to the lack of distractions.36Broking remote lab
A suggestion was to set up a business model providing remote labs based around high quality experiments through a broker.37Customers would range from universities and individuals to industry. The broker would be responsible for overall management (quality control, handling of payments, allocating time slots and real time monitoring of the labs) and marketing of the remote labs sourced from a number of institutions. Some potential problems include a customer’s lack of ability or willingness to pay for this service, and the availability of sufficient high quality labs.Formative assessments
There is limited evidence of formative assessments being used in e-labs.38 Embedded assessments in remote labs were thus successfully applied at the universities of Colorado and Houston for five labs ranging from free space propagation and material dispersion in optical fibers to optical detector characterization.Remote labs on their own are inadequate for accredditation
It was maintained that the most serious deficiency in an online program is meeting the ABET requirements for meaningful lab exercises.39 This can be effectively dealt with by having short duration face-to-face, high intensity laboratory (boot) camps.Sustainability and maintenance of remote labs
The majority of remote and virtual labs have failed beyond the development phase.40 The main problem with the sustainability of these labs would appear to be the lack of a business model after the development phase and the absence of 24/7 access on a worldwide basis. These labs should be as versatile as possible to cater for a wide range of participants, ranging from students and graduates to engineers and technicians, and the learning objects provided should be able to be reconfigured by the educator for her target audience (perhaps based around an LMS).
Suggestions for improvement to the sustainability are to replace aging hardware and software. In order to make instrument replacement as easy as possible, software should be configured (and programmed) at the highest possible level. The graphical user interface should have extensive functionality, be as interactive as possible, be easy to use and provide realistic visualization of the experimental environment.
In order to maintain a high degree of reliability, periodic maintenance on both software and hardware should be performed, focusing especially on components such as computer power supplies, batteries and switches and performing calibration where required. Software errors should be logged and investigated.University of technology
A qualitative survey revealed that students at University of Technology Sydney undertaking two remote labs, Programmable Logic Controllers and the Water Level Dynamics, considered that increased time and ease of recordkeeping (added to the natural one of convenience) were positive traits of remote labs.41 There were negative perceptions about problems with the interface (the poorly configured desktop and video lag) and isolation from others (and no ability to get immediate feedback to questions).Scalability of remote labs
One of the concerns with remote labs is the issue of scalability in terms of varying numbers of experiments and clients wanting to access them.42 At the University of Houston, a proposed solution to this problem was via a scheduler web server that was used to interface to the client user providing her with a customized direct link to the experiment. Each experiment included a web service to provide the data and the interface to the client. All management of users and authorization was done through the scheduler web server.Pedagogical implications of remote labs
While there is a strong focus on the technology of remote labs, there has been very little emphasis on the pedagogical implications of the technology.43 The University of South Queensl and created a learning activity for primary school students (13 aged between seven and 12 years) based on their Remote Access Laboratories (RAL) entitled Robot RAL-ly. This was based around remotely controlled track-based robots with webcam monitoring and an arena in which to stage the event. A race course had to be designed by each team for their robotic equipment. They then had to remotely navigate their way around the other team’s track. Several themes of autonomy and motivation emerged: collaboration and teamwork, communication, problem solving, learning transfer and reflection and metacognitive awareness (of the skills and learning built up).Remote controlled power switch for cycling power
It is possible to incorporate a remote-controlled power switch (for the lab equipment and PCs) through the PSTN telephone network (by simply dialing up) to perform a power cycle when the system locks up.44Problems with JAVA
The most commonly used approach for control and monitoring of remote devices and labs has been the Java Applet.45However, this has been proved difficult for applications where the IT administrator doesn’t allow for installation of additional software. Java and Flash take time to load up, and use of the Applet on non-Windows operating systems can be problematic.
The following operation is followed with LabVIEW from instrument to the client interface:
• The data acquisition board sends its data to LabVIEW.
• LabVIEW writes its data to the MySQL database and flags it as ready.
• The PHP script reads the values from the database and returns it through XML to the client interface.Ease of use is vital
It is critical during the use of remote labs that you make it as easy as possible to use.46 This is especially for an impatient audience (as many students are) who want to perform their experiments with minimum queuing to use the experiment and little to no downtime with the remote equipment. Having the real lab remotely located with all the attendant communication problems can make the whole process unworkable and you may need to consider simply using less realistic but logically easier local simulation experiments.
11.5 Converting from traditional labs
As noted earlier, there are various choices when converting a classic lab-based course to a distance learning one. This includes removing the exercises from the course, replacing labs with virtual labs, providing each remote student with a lab kit or having a mobile lab that can be taken to each remote site.47 Another option is to use remote labs.
Careful measured consideration should be given to whether a traditional lab experience can be converted to a remote lab or not, based on optimizing the learning experience.48 The Fluid Mechanics course unit at Curtin University was used to demonstrate the factors that need to be considered in converting a traditional lab to a remote one. These included such elements as:Visalization of flow patterns.It was vital for a student to see the laminar and turbulent flow patterns. With appropriate webcams, it is possible for the students to have a better viewing experience remotely than in a crowded lab.
Lab scheduling.In the fluid mechanics lab, there were over 500 students who had five weeks in which to undertake the lab on one piece of equipment. Remote access on a 24/7 basis widened the participation possibilities enormously.
Participation within the lab itself.With this large number of students, they were broken into teams of tens or more. This reduced the individual student's learning experience and a remote lab could significantly improve an individual's time in working on an experiment.
Reapeating the lab session.It is only once the student has commenced analyzing results that they may need to repeat the experiment to gather missing data. Under the current scenario this was only possible with remote labs.
Satellite campuses.Curtin had another satellite campus at Miri in Malaysia. Engineering students at this campus would be able to access this lab remotely and thus obviate the need to build another residential lab on the Malaysian campus.
Variability of physical parameters.There was concern that students would misinterpret the dynamically varying measurements of flowrate and ascribe these as faulty. Alternatively, if the software averaged out the changing parameters, these would not be interpreted correctly by a remote student.
Remote measurement and control.It would be easy to digitize all fluid measurement parameters and to perform control of the pump's speed on a remote basis.
Missing sensory information and true tactile hands-on activity. There is often some disconnect with remote labs in terms of feeding back appropriate stimuli to all the senses of the remote student observer. For example, the smell of hot oil and the sounds of fluid running through a pipe are a key part of the experiment and would be missing from the experience. The tactile sensation of adjusting the stopcock to allow more fluid to flow would also be missing (and be replaced by a simple setpoint change on a computer).
Overall, in consideration of all the factors above, on balance it was thought appropriate to convert this lab into a remote lab with a positive outcome for a student. These factors could be aggregated into the three categories: learning (due to more individual access), equipment (due to no specific skill required with the lab hardware) and cohort factors (large number of students with low level of access).
11.6 General requirements There are generally two approaches to people working on remote experiments; either as one person or as a team where one controls the experiment and the others merely monitor.49
There are three different profiles for working with remote labs.50 These are:
• The students should be able to book a lab for a specified time, undertake the lab, upload or download data and communicate with each other.
• The instructor should be able to configure, modify or set the parameters for the labs (e.g. duration), assess the student results, communicate with students both synchronously or asynchronously.
• The administrator should have the same privileges as the instructor but in addition should be able to configure new users with their passwords.
Some suggestions for participants working in the labs were:51
• A strong theoretical background on the subject matter is necessary.
• The training phase must be carefully planned with collaboration with other learners arranged.
• Interactivity must be made as high as possible with extensive control and data acquisition of equipment.
• Interactivity•Learners must be thoroughly assessed at the conclusion of the work.The remote lab assistant
It is ideal to ensure the only interaction is between the remote participants and lab apparatus, but sometimes a lab assistant will be required to undertake the following tasks:52
• Correcting of malfunctions.
• Resetting equipment and computers.
• Assisting remote users.
• Tripping breakers.
• Replacing parts.
• Moving cameras/microphones.
• Altering connections.
• Cleaning up and preparing new experiments.Design issues for remote labs
The following issues are important to consider when designing remote labs:53
• Modularity, expandability and scalability.
• Easy interfacing to existing communication standards.
• Computer platform independence.
Thorough testing of the complete system before release is necessary, mainly due to its complexity.54 In addition, registration and booking for lab resources is critical to handle the inevitable peaks in dem and by students (doing their tasks immediately before submission of their lab exercises). For those more complex labs, with multiple students working together, it is important that the software supports a collaborative approach.
Finally, real time access to a lab is regarded as far more useful than a queued batching (and off-line) approach, especially for individuals being exposed to the technology for the first time.
• Integration of different items of data such as real time process data, interaction with other students and multimedia data.
• Achievement of Quality of Service parameters.
• Multi-platform, easy to install, secure and cheap (or free) student client software.
• The ability to download the data gathered for later analysis.
• Simultaneous access for multiple players such as instructors, administrators and students.
• Easy management of students, resources and assessments of experiments.
• Reliable software. Exacerbated by the remoteness, any problems will irritate students.
• The ability to be powercycled remotely to recover from crashes and to be placed into a defined initial state.
• A remote lab that is a clearly defined and required part of the course and offer value to the student.
• The lab must offer a wide variety of experimentation, thus allowing the bright students considerable latitude in experimentation; a simple cookbook lab with limited scope for experimentation lab will bore them.
• Ease of use.
• Portability of all equipment.
• Control of remote objects.
• Ability to collect large amounts of data from the remote site.
• Ability to archive all the interactions.
• True interactivity between local and remote sites.
• Ability for students to access all PCs in the lab.
• Ability to manage all lab resources easily and effectively
• Simple remote access.
• Robust equipment.
• Responsive test equipment (with minimal delays).
• Extendable practical exercises.58
• A simple, effective and easy-to-grasp user interface.
• No additional hardware or software required for the client or student machine.
• 24/7 availability.
• Minimal or no interaction and support required from on-site lab personnel.
• No perceivable delays in the operation of the lab.
• An interesting and attractive experience for students (with audio and video support to improve the experience).
• A built-in computer interface (via Ethernet or USB).
• Electronically-based lab sensors.
• Video and web tutorials to explain the key attributes of the remote lab.
• Secure client, server and remote lab equipment.
• The ability for students to actively participate in the lab experiments with decisions that will impact on the final results.
• Experimental results that are subjected to analysis through proof-testing, comparison to standards (including degree of error).
The more immersive the experience in a remote lab, the more likely it is to be considered by a student to be useful.59
The student interaction with a remote lab should be as close as possible to that with an actual lab.60
For maximum impact (not absolutely critical), the dynamic behavior has to be visually observable. So temperature measurement may not be as powerful as motion control, for example.61
The dynamic behavior has to be (as Goldilocks is wont to say) “just right”. Neither too fast (otherwise this is lost on the slower internet-based video) or too slow (where the interactivity is lost).
The entire system has to easily and automatically revert back to an initial state ready for the next student. Having a full-time technician present all the time is simply not economical.
A good design concept to focus on is based around authentic learning environments.Authentic learning environments as applied to remote physics labs
An authentic learning environment (as discussed in Chapter 2) was established for the remote physics labs at the Berlin Institute of Technology for the following reasons:62
• The context of the work is authentic as the remote labs are authentic to a future career as engineers.
• It comprises an extended set of experimental, analysis and report writing tasks to be performed over several days.
• Students examine the remote lab from a number of different perspectives such as details of the experiment, videos, simulations, remote experiments, wikis and discussion forums.
• The students collaborate in groups of three to undertake the experiment, analyze the data and write the report.
• The students have choices in what experiments they undertake, how they structure their groups and what time line they use to perform the experiment.
• The work is performed across different subject areas and requires skills ranging from scientific to analysis and report writing.
• The final product (the report) represents the culmination of the work and is a polished product in its own right.
• A range of competing solutions and diversity in outcomes could be seen by the method of undertaking the experiments, structuring the groups and writing about the results of the experiment.
• The assessment was seamlessly built into the remote labs and was built around ability to perform the experiments, analyze the results and prepare the report.
• Ill-defined activities were not obvious but could be seen by the way the students structured their groups and wrote their reports.
11.7 Software requirements
It has been suggested that software aspects tend to be neglected in the design of remote labs.63 Client software can be classified into two main categories: desktop (applications that run on the computer) and web-based (browser-based). Desktop applications (incl. C++, C#, Java, .NET, Python) are generally more powerful than browser-based ones and can result in 3D graphics generation (something which would be difficult to achieve with web applications). Note the golden rule: the more powerful a technology, the less universal it is. However, web-based applications do provide more portability and less intrusiveness (access to hard disk, read any file on a computer). The main challenge with increased intrusiveness is compromising the security of the client’s machine.
The challenge in using non-http based technologies is the need to reconfigure a firewall to allow communications to the remote lab from the server. Hence, web services are often preferred but the drawback here is performance; they are normally slower than proprietary technologies.
Certain characteristics were listed as critical by the University of Deusto in assessing client applications for the optimal technology. These included:Universality: Ease of access of the client with minimal restrictions. This includes such elements as cross platform (operating on different operating systems); accessibility (for disabled people) and compatibility with web browsers.
Security/Standards: A high level of security based around secure standards. This includes such elements as intrusiveness (e.g. permission to access hard disk or establish a connection); standardization, installation required (e.g. software installation) and network protocols available.
Power: The level of power of the client.
This includes such elements as audio/video, bandwidth efficiency, flexibility for different contexts and ability to be used on different mobile platforms.Development work: How easy and effective the tools provided can be. This includes such elements as development tools, price, independence of developers and users from provider and the size of the community of developers.
According to the survey, AJAX was the technology of choice. However, its main disadvantage was lack of audio and high quality video (which could be provided by Adobe Flash or Java).
Of three server technologies (Java, .NET and Python), Python was considered because it is a powerful dynamically typed language with strong open source support, it is fast to develop and large blue-chip companies use it.Use of web services in a remote lab
This approach with the use of web services was applied to an undergraduate course in solid state electronic circuits with the remote lab-based on tests (such as extraction of the threshold voltage parameter of the MOSFET) on transistors fabricated as part of the course.Use cases
A collection of use cases (courtesy of Rainer Bartz and Daniel Cox’s comprehensive work in this regard) for remote labs include:65
Registration of a User with a Remote Lab. Before access is provided, key elements of information are required from a user (before being formally approved).
Log in Process to Remote Lab. A log in procedure is required before the user can commence using the remote lab.
Management of Time Slots. Users prefer not to be kept waiting for access to the remote lab, which is a scarce resource, and hence some management procedure for guaranteeing a user access at some mutually negotiated time is required.
Selection of Lab Model and Experiments. When a user commences work, she should be able to access both a specific model and experiment (of which many different types could run on one model). Only one experiment can naturally be run on a model at a specific time.
Configuration and Execution of an Experiment. Typical configuration parameters for an experiment include duration of experiment, sample times, filtering, and PID constants.
Retrieval of Experimental Results. The main reason for running an experiment is to be able to access authentic results through either a push (all results sent to user once experiment concluded) or pull (user has an interactive way of viewing the results).
Change of User Information. Users should be able to modify their parameters and also retrieve details of old records (including handling the issues of loss of password).
Management of Configuration and Result Data. Old configuration parameters of an experiment should be saved, thus saving a user from re-entering these data values for every experiment attempted. In addition, all data from previous experiments should be saved for later retrieval (including means for archiving it or deleting it eventually).
Administration of a Remote Lab. Typical tasks in this respect would include such elements as (dis)approving a new registration, disabling an existing user, provision of data to users, modification of time slots, disabling of current experiments or models. Finally, management statistics on usage of remote labs and users are also useful to review.Entities and their attributes
The main entities (courtesy of Rainer Bartz and Daniel Cox) are as follows:
• User Entity. UserID / LastName / FirstName / UserEmail / LoginName / Password / Active / Admin / RegisteredOn / Purpose / Salutation / Organization / OtherInfo.
• Admin refers to administration rights provided (or not). Active indicates whether the user is actively involved working on the remote lab at the time.
• Model Entity (the Physical Equipment). ModelID / ModelName / ModelURL / Active.
• Experiment Entity. ExperimentID / ExperimentName / ModelI D / ResultsTableName / ParameterNames / ParameterUnits / ExperimentURL / Active. Any number of experiments can be defined for a specific model. The ParameterNames and ParameterUnits comprise comma-separated strings. The Experiment entity is activated (and made visible to students) when the administrators have confirmed that it is working fine and should be added to the list of active experiments.
• LoginInfo Entity LoginID / UserID / LoginStart / LoginEnd. The LoginID is purely for internal indexing purposes.
• ReservationInfo Entity. ReservationID / UserID / ModelID / StartYear / StartMonth / StartDay / StartHour / StartMinute This stores the reservation details for each user. It is obviously crucial to avoid having two users with the same time slots.
• SessionInfo Entity. SessionID / UserID / ModelID / ExperimentID / SessionStart / SessionEnd / Active / DataExist. This records a session that a user engages in. The DataExist indicates that there are data resulting from the experiment.Design of graphical user interfaces
Although the focus in remote labs has often been on the technical aspects of interfacing to the equipment and making the lab available to a remote user, the graphical user interface (GUI) must surely be the most critical feature of a remote lab as this is the link between user and lab.66 One major limitation of remote labs is the poor interface design. Many complaints relating the use of remote labs have been related to the difficult or buggy operation of the graphical user interface.
More emphasis should be placed on proper user-centered interface design.67 Having realistic looking interfaces (similar to the real-world lab) should be the focus rather than the usual alternative of providing circuit schematics. An example with testing an op amp would be to provide a realistic picture of the chip and the resistors with switches (for connecting them in and out) on a circuit board rather than an electronic circuit schematic of the op amp and resistors.
A few suggestions for good design of graphical user interfaces include:
• It should be intuitive and idiot-proof (or as we often joke, orangutan-proof) in its construction. It should not distract the user from the actual experiment and equipment and should effectively be transparent. The KISS (Keep it Simple and Stupid) principle should be kept in mind.
• It should be as authentic as possible and reflect real equipment used. This may contradict the previous comment; as some equipment operation can be rather obtuse and most are rapidly migrating to PC-based software interfaces, anyway.
• As in a traditional lab, it should allow for multiple users working together in a collaborative environment.
• It should run on a standard computer connected to the internet with minimal software add-ons (preferably within a browser), all of which can be accessed freely on the web.
Other important design principles, which also relate to the overall design of the system, are:68
• The GUI should reflect the status of the remote lab equipment as quickly as possible without any irritating delays (including confirmation of the control actions performed by the user).
• All terminology used should be in simple user language rather than arcane, jargon that’s difficult to understand.
• There should be a quick exit from unwanted selections. When the user makes a mistake with an erroneous selection, she should be able to exit quickly without having to engage with a long drawn and often incomprehensible dialogue.
• The entire interface should be consistent and maintain standards throughout in the use of language, status and control actions.
• Actively assist the user in avoiding errors. A good design will prevent the user making errors and perhaps ask for user confirmation for a particular action if there is perhaps some unavoidable ambiguity.
• Avoid requiring the user to remember particular steps in using the GUI. All actions and interpretations of the GUI should be easy to underst and or instructions on how to use a particular feature easy to retrieve. One should avoid requiring the user to memorize particular steps.
• Accelerate the expert user to her destination. The experienced expert with the system should be able to quickly learn how to take short cuts, especially with tasks that are repetitively done that would be transparent to the novice.
• Treat every word in the dialogues as gold. Avoid overwhelming users with irrelevant words that can be confusing and are not as easy to interpret as objects. The fewer words used, the better; every word should be meaningful and simple. Bear in mind that many users will not have English as their first language.
• Recover gracefully, quickly and knowledgeably from errors. All error messages should be easily understood (and avoid error codes) showing the user where she has gone wrong and how to rectify the problem expeditiously.
• Documentation and help dialogues are normally minimized in a well-developed GUI. The system should preferably be used without any documentation. In the worse case scenario where it is required, it should be focused on the problem in hand and express a solution in simple English.
11.8 Overall design considerations
A designer of a remote lab should consider the following issues:69Type of experiment
• How will the architecture impact on the experiment?
• Can the experiment be shared simultaneously with multiple users (i.e. time division multiplexing)?
• If no simultaneous sharing is possible, how much time does each user require and how will queuing affect her?
• Are video streams required and if so, at what resolution and frame rate?Scalability
• How does the architecture respond with peaks in users?
• What is the limit?
• Is the scaling horizontal (more nodes meaning more connections) or vertical (more memory and CPU to one node means more connections)?Maintainability
• Can the remote lab be integrated easily into the IT services of other universities?
• Is there an advanced user management system in place with multiple types of users (e.g. instructors, lab administrators, users, system administrators)?
• Is there a detailed statistical log provided of users?Security
• Is security a key part of the design?
• Are secure communication protocols supported (e.g. to avoid code injection such as SQL/LDAP/XParth injection)?
• What security policies have been developed for the remote lab development?Dependence on protocol
• Does the architecture require a specific topology and type of protocol?
• Are the different experimental kits and application servers required to be in the same physical location?
• Are multiple protocols supported for different requirements (e.g. binary for real time interaction with a device)?
• Do the protocols change based on the security requirements?Service orientated architecture
• Is the architecture based around a service-oriented architecture?
• Is the remote lab deployed as a service (such as SOAP) or is it for purely browser-based clients?
• Can other services be built upon the top of remote labs?11.9 Jump starting remote labs
Why the slowness in migrating to online labs?
A few reasons are blamed for the tardy movement across to online labs.70 These include the time it takes to convert materials into online format and necessary training of staff, transfer credits for online lab courses being unacceptable to other institutions, faculty resistance, reluctance to embrace the new online instructor role (as a “guide on the side” rather than a “sage on the stage”), costs to change over, doubts about whether online is a better approach to the traditional way, the ongoing investment in time to present successfully online and finally, liability for the student performing labs remotely.
The following are other barriers to adoption of remote labs:71
• Poor understanding of what remote labs are and how they can add value to engineering learning.
• Poor understanding of pedagogy in the remote lab environment.
• The transformation of learning tasks from those for a classical to a remote lab.
• The necessity of more learning activity development in the remote lab area.
• Technical expertise that needs to be developed in constructing and conducting remote lab experiments.
• Extensive time and effort required in this new domain.Popularizing remote labs
The main issues that have to be dealt with effectively to popularize remote labs include:72
• The design of remote labs is an amalgam of a number of different engineering and computing disciplines such as data acquisition, networking, web security and mechanical and electrical engineering.
• The design of remote labs needs to be made more modular with a common framework to provide quicker and easier design and construction.
• Commercial off the shelf products (COTS) are needed to accelerate the improvement of the design and development of remote labs.
• As Learning Management Systems are a key element in online learning, they need to be more closely integrated with remote labs.
• Maintenance of remote labs is an unusual skill and more effort needs to be put into training staff to be able to support this technology.
• There is a low awareness by academic administrative staff in this technology and thus there is no support follow-through once a remote lab has been constructed.
• Finally, industry needs to embrace this technology more widely, as it can be enormously useful for remote support, training and remote diagnostics.
The recommendation was made to have a common integrated framework (especially for indexing facilities, unique log ins, file sharing and seamless access) in order to have a global network of remote labs.73Concerns about underutilisation of remote labs
Remote labs have been around at least since 1996 with a rapidly increasing number being launched.74 Although there is admittedly no hard evidence available, there is anecdotal evidence that remote labs are underutilized (and in addition, represent an unrealized opportunity). For remHHHote labs to be adopted, they need to be perceived as useful and easy to use and there needs to be peer enthusiasm for the technology. A few suggestions to increase the take-up of remote labs included providing support to academics in the design process, helping align the labs with learning objectives, training in their use and interaction with others who have successfully installed the remote labs. Finally, there should be provision of sufficient time and resources for academics to study and apply the technology.
Finally, the painful observation is made that despite their being over a decade of work done on them, most remote labs end up being shut down after their funding dries up.75Commercial software offerings
One of the surprising things has been the lack of much commercial software development for remote labs. Electromeet, put together by the authors, is one commercial offering but other than this, there is not much else around. However, one notable example of a commercial lab offering is NETLAB+.76
The Net Development Group has developed the remote lab system NETLAB+ with a focus on Cisco Networking Academy program. Colleges that have been using the NETLAB+ appliance have been Stanly Community College (North Carolina) and Honolulu Community College (Hawaii) for their computing, networking technology and information systems security programs offered completely in an online format for two-year degree programs. NETLAB+ also provides support for virtual machines, being a remote PC that operates on virtualized hardware (from such companies as VMWare).Large scale of deployment of remote labs
Little research has been conducted in the area of large-scale deployment of remote and mobile labs especially as far as infrastructure and network security.77
The following are considered minimal requirements for large-scale remote laboratory deployment:78
• Simple user interfaces.
• Standard internet protocols.
• High level of computer and internet security for all components.
• High level of scalability.
• Remote lab management functionality.
• Sustainability.11.10 Conclusion
Online labs are an important addition to online learning to provide easier access to the wider community, improved active learning and allowing for more learning via discovery through experimentation and associated trial and error, for example.79
Remote labs are acknowledged as providing the affordable real world experience.80 It is possible that if the interest in remote labs grows dramatically, there will be a vibrant industry of provision of remote labs located throughout the world, perhaps brokered by private companies.81 There are already examples of low-cost remote labs available for rental in the IT training and certification area (e.g. configuration of routers). One suggestion is that distance learners using remote labs as opposed to attending compressed lab sessions intermittently on campus may be able to bond more closely when working on the web in an asynchronous community and thus to form a community of learners.82
There are nevertheless some doubts about whether remote labs would ever be able to replace hands-on exercises in a real laboratory, but the hope is that in providing a holistic training experience, remote labs will be combined with virtual and the conventional face-to-face ones.83
Rather than having a one-dimensional approach, ultimately, a blended approach with a combination of remote labs, simulations and classical labs is probably the optimum solution. For example, students could begin the lab work in a conventional lab and then continue the work after hours by accessing the same lab remotely.
It should be noted that remote labs can’t replace all face-to-face lab and field training.84 For example, troubleshooting strange malfunctions of equipment often requires on-the-job training with an experienced mentor. Most students still indicate that they would prefer conventional labs providing real hands-on experiences as they feel they learn more from them.85
It is likely that a mixed or hybrid approach of using remote labs as a supplement to the normal residential labs is the way to go for an undergraduate experience for someone (as a 17-year-old) who has not worked with the nitty-gritty of engineering in practice. Another area where remote labs have an unassailable advantage (even for 17-year-old undergraduate engineering students) is in developing software for an expensive development tool for digital signal processing. Microprocessor development tools can change frequently and can be very expensive. At the end of the day, a remote lab is considerably better than watching a video of a lab if this is all that is available.86
The hybrid approach offers an opportunity for an excellent learning experience. Initially, the basic theory is reviewed, then a simulator is used to illustrate the basic principles and the session is concluded with a hands-on real lab, but this approach should be restricted to students who have built up a solid degree of experience in working with real components and equipment (as opposed to those who have never worked in a real lab before).87
One of the surprises with the growth of remote labs is the lack of connectedness to other online learning resources (such as a complete course with a series of remote labs forming an integral part).88
Finally, in constructing remote labs one should not forget an outstanding interface and interactivity between the instruments and the web conferencing software and the instructors and learners. And naturally, as with the classroom-based training, the course materials and instruction should be of the highest quality.89
Key points and applications
The following are the key points and applications from this chapter entitled: Remote Laboratory Approaches.
1. Learners find a well-constructed remote lab with a simple interface useful and equivalent to that of a traditional lab, but there are misgivings about the lack of authentic hands-on work.
2. A typical architecture comprises a learning management system through which learners access a series of lab servers which allow access to the actual lab equipment.
3. There are five main components of a remote lab:
• Web and database server containing user accounts, experimental documentation and monitor of students’ experimental activities.
• Collaborative server allowing for a synchronous web conferencing experience.
• Experimental computer connected to switching matrix.
• Switching matrix allowing for remote switching of experimental components.
• Experimental equipment.
4. Successful remote access technologies include Teamviewer, GoToMyPC and LogMeIn.
5. Some elements required to build a successful remote lab include:
• The need to avoid errors in the operation of the labs due to the distance and possibly multiple users accessing a lab.
• System response time of only a few seconds (preferably a maximum of 150 milliseconds).
• Time scheduling systems for multiple students trying to access the same lab.
• Importance of Reality with easy-to-use graphical user interface.
• Good lighting.
• Well-trained instructors.
• Lab work counting towards the final result of the course.
• Encouragement for frequent use of the labs
6. Some design principles for the Graphical User Interface (GUI) for remote labs include:
• Keep it simple and intuitive.
• Authenticity should underpin everything.
• Multiple users should be able to work together.
• The student should be able to interface with a standard computer with minimal software add-ons (preferably within the browser).
• The GUI should have no delays displaying key information.
• All terminology used should be in simple English.
• There should be a quick exit provided from unwanted menu selections.
• Interfaces should be consistent and standardised.
• Avoid requiring user to remember particular steps with the GUI.
• Error messages should be easily understood.
7. Overall design considerations:
• Type of experiment (how will the architecture impact on the experiment; multiple users and queuing issues).
• Scalability (how will system cope with peak number of users).
• Maintainability (integrated easily into overall IT services).
• Security (is this a key part of the design, and are secure communications protocols supported?).
• Dependence on protocols (architecture requiring a specific protocol).