THD204 - TMA06 - This is the critique of the Group 'Spritely Comes Home'

Summary

This Group 1 report has been prepared as the basis for discussion for our client Acme Robotics Inc. to investigate the proposed manufacture of, and future development of, SPRITELY robotic aids for the disabled. In addition to analysing specific theme / design issues of the SPRITELY system, the principles of revised system development as a whole will be outlined. The report will also outline the collaboration of the OPUS Foundation and the Open University, and the technological and social issues surrounding future development. The information contained with the report is derived from a number of technology-based publications e.g. and from trusted Internet sites, and therefore presents an accurate assessment of the principles embodied in the report. There are a number of different factors that need to be taken into account. Some of the principle considerations are that SPRITELY robot in its present form: -

· May not be safe enough to be permitted free-range access within the house;

· Does not provide the disabled user with sufficient degree of independence;

· Performs limited actions, derived from sound orientated commands that may not be useful or meaningful to the disabled user.

It is not implied that the existing command control should in any way be reduced but that the detailed level of control should be examined e.g. via voice command, to reduce the possibility of error, and to improve the value of the information produced by the robot. The report will concentrate on four major themes / Issues for discussion prioritised as follows: -

1. Safety and Security - The Safety and Security of the disabled person is paramount and cannot and must not be compromised.

2. Access and Inequality - Access to new IT developments such as SPRITELY are not distributed evenly.

3. Human Technology Interaction (HTI) - A wide range and variety of input and output devices that are used to translate information will need to be analysed so that they are adapted suitably for the task in hand.

4. Control -The use of SPRITELY may give the user a greater ability of self-determination and should ultimately lead to a return (or introduction to) the workplace.

The rationale for the prioritisation process was influenced by the following factors and assumes:

• · the Open University and OPUS utilise the right development environment;
• · the re-development program is undertaken in a logical and systematic manner;
• · the nature of the finished product, in terms of parameters and uses of the model, is appropriate for the disabled when they attempt to use it.

These matters are investigated more closely within the report as they will determine the basic decision to use a particular model, its implementation and the general involvement of individuals in the framework of the project development, disabled use and post-use reactions.

Other theme and design issues e.g. Pace of Change, Technological Determinism and Standards and Compatibility are not discussed implicitly due to time constraints, limited resources, and the minimum information supplied regarding the launch of SPRITELY. These design issues and others may be explored to some degree within the specific report areas. The recommendations for continuing further investment in manufacturing costs and future development of SPRITELY are:

1. The price of SPRITELY could fall to more affordable levels subject to increases in demand;
2. Modular development providing support to a cross-range of disabilities;
3. The well established standards for object orientated programming tools;
4. The enhanced quality of life as experienced by the disabled user.

In short it is suggested that SPRITELY should become a modular development to provide custom made mobile robot facilities which will have significant benefits for the disabled. It will enable them to enjoy an enhanced way of life, and gain awareness of modern technological systems. Future project development will help to actively promote Acme Robotic Inc. in a professional manner. This will create a favourable impression on the disabled users, the local community, and other related business areas. As a potential employer Acme Robotic Inc. will value Open University students' and OPUS research, and the development experience with new technology.

SPRITELY - Safety and Security Critique

1.1 Introduction

This part of the report examines certain Safety and Security aspects of SPRITELY raised by the article. Concerns discussed included claims of testing the safety of the software and control and guidance systems. Safety is paramount in ensuring the security of user and robot alike.

1.2 "At present SPRITELY is undergoing final testing with the Health and Safety Executive"

The H&SE do not test appliances before public release, their remit is to enforce the Health and Safety at Work Act and issue directives with regard to H&S at work. HSE mission statement "To ensure that risks to people's health and safety from work activities are properly controlled"[1]. Their role is primarily in the field of occupational health. Although they do have their own R&D and testing departments [2] these are not concerned with manufactured products unless an injury or death has been caused.

1.3 "A cordless DECT phone provides access to the emergency services"

The DECT phone requires the use of a base station connected to the phone socket. My concerns over the use of this particular method are:

• 1). In the event of a power failure the phone is useless as the base station will be off and it clearly states in the BT manual for our phones at home that "In the event of a power failure alternative means of communication will be required" [3].
• 2). Who is going to charge the phone - is this a function of SPRITELY or is the user expected to do this - once the phone has lost charge it is useless for at least an hour (personal experience).
• 3). Most important of all, will the DECT phone interfere with the guidance and control systems.

In mitigation the use of a DECT phone would be extremely useful to someone engaged in teleworking.

1.4 "The base model is acoustically controlled with a stored repertoire of 100 sounds"

1) While it might work adequately in a laboratory environment where the background noise is controlled how is it going to cope with a normal domestic situation where the DECT phone, TV, radio, stereo or other appliances such as Hoovers and washing machines are in operation. Under certain circumstances a sound from one of these sources could make the robot go off and perform a task unexpectedly and dependant on the task could place the disabled person in jeopardy. I use IBM's Via Voice dictation package and although it has learnt my voice pattern it still confuses certain words (personal experience).

2) What commands is it expecting - whatever it expects the user is going to have to learn the 'command set' to be able to use this robot so there is an issue of user training. If the user is not properly trained the robot could be of little use or could be dangerous.

"users rarely had adequate training in the use of the new equipment, which often led to the equipment not being utilised optimally"[4].

1.5 "SPRITELY's versatility arises from the use of object orientated programming tools and neural networks for the command and guidance system."

Whether it uses object orientated or conventional programming is immaterial; the robot software must be 100% safe ensuring security for both itself and user. People always assume that the results of programming are always correct, in this case an error in the program could injure the very person the robot is supposed to assist. "A common mistake in engineering, in this case and many others, is to put too much confidence in software" [5].

SPRITELY's software is 'safety critical' and as such should be tested by a third party. Testing by the design team is the normal way of development, however they become used to idiosyncrasies of the software and so become complacent with its foibles. This closeness to the design means that potential errors can be missed, and testing is usually performed in a controlled environment. Because of this the robot should be tested by people who have no connection with the project to ensure thorough rigorous unbiased testing. "There is a need for an independent validation agency: It is impossible to test software completely and difficult to test one's own design in an unbiased way"[6].

Having made such a point about the software there remains the possibility of faults whether electronic/electrical/mechanical or the software itself causing the robot to become unsafe in operation. Therefore in a complex system like SPRITELY the software must be fault tolerant, since it could be the hardware that is at fault and must "fail-safe". That is to say that the systems either shuts down completely or the offending process is terminated pending maintenance the safest option is always taken.

"Software failures may be caused by such undetected hardware errors as transient faults causing mutilation of data, security violations, human mistakes during operation and maintenance, errors in underlying or supporting software, or interfacing problems with other parts of the system such as timing errors"[7].

1.6 "self-learning capabilities essential for 'structured environments' such as the home or office."

SPRITELY we are led to believe, can learn its way around any home or office. However the average home is anything but a 'structured environment', people and pets move about, doors open and close, the furniture gets moved things are left lying on the floor etc. The very fact that this is an 'unstructured environment' will probably cause the robot to become easily lost. The article also says the robot has a guidance system but in this 'unstructured environment' will it be of any use since objects are frequently being moved thus confusing the guidance system hence the robot will be spending most of its time relearning it's way round.

This series of short extracts from the New Scientist article entitled 'Jobs for Bots' [8] concerning a robot called Cye sums up the problems. "If it hits and obstacle it often gets stuck", "If you rearrange the room Cye has to be re-educated" and "Cye loses it position by 2.5cm approx. in every metre due to wheel slippage". Therefore the robot in its current form will be unsuitable for the vast majority of dwellings, as they will be of the 'unstructured environment' type.

The article also omits to tell us if it has a collision avoidance system since when it is lost it will have to re-learn its position and will probably be bumping into objects. The collision avoidance system would stop the robot crashing into things or the disabled person, however no reference to a collision avoidance system is made. Since we are dealing with physically disabled people who most likely have limited mobility they will not be able to get out of the way of the robot thus risking injury. This is most unsatisfactory because the safety of these people is the number one priority. "If safety is defined in terms of a mishap or catastrophic event, then difficulties arise from the fact that mishaps are often multifactorial and may involve conditions in the environment (i.e., not part of the system being considered or evaluated) over which the designer has no control"[9].

1.7 Security of SPRITELY

It is not only the client that is at risk the robot itself could also be at risk from children or pets; we used to have a dog that attacked the Hoover! (personal experience). So there is potential danger from this area, which could have serious repercussions, should humans or pets be injured.

"As robots join family life they bring new dangers with them, for owners and manufacturers alike. Just one injured toddler or electrocuted dog would destroy the market forever"[10].

1.8 Safety and Security Conclusions

SPRITELY in its current form is not yet ready to be let lose on humans let alone the disabled. Since SPRITELY is to be used in homes of the disabled it is vital that the software is error free. The software must be demonstrated as being totally safe before any trials. The guidance system requires work to bring it up to a standard that can cope with the average house. The system must be 100% safe incorporating "fail-safe" mechanisms to protect human and robot alike. "An integration of these safety aspects indicates that the most important ingredient surrounding the integrity of robotic systems that could be employed for the handicapped and elderly is a 'guaranteed 100 % assurance that the robot will operate according to its designed characteristics" [11].

These are not insurmountable problems and SPRITELY should make a good development platform for future robotic aids.

Group Conclusion

The scope and the parameters of the SPRITELY product have been determined through collaboration between the Open University, the OPUS Foundation, and its intended purchasers or users. ACME Robotics Inc. needs to be aware of the fact that customers may be over-optimistic about what they think SPRITELY can do. Clearly it is better for the disabled user to identify the potential differences between what they need and what they think they are getting in the early stages of re-development.

Having modular architecture should allow ACME Robotics Inc. to argue for its implementation on the basis that it closely matches the requirements of many disabled customers, and that it will, consequently, have longer-term benefits to the company as a whole. The security and safety, access, human-technology interaction and control issues will relate to AMCE identifying the fact that the option will offer the opportunity for the company to reduce the failure rate of SPRITELY, or provide a general improvement in the quality of service. Again this might mean that ACME will be able to enjoy certain financial benefits if the number of faults and mistakes are minimised.

The report has assessed the performance of SPRITELY in terms of its potential and suggests a number of baseline changes that ACME should consider before development of its manufacturing plans.

1. Safety & Security -SPRITLEY is to be used with disabled people, therefore there cannot and must not be any compromise where their safety is at stake. There are many concerns over the Safety and Security of SPRITELY particularly with regard claims made about the testing. ACME Robotics expertise in the robotics field and their knowledge of current regulations regarding the use of robots will be complementary to OPUS and the OU in finding the appropriate solutions for those concerns. ACME Robotics will be able to move the design forward to a model capable of providing rehabilitation assistance to the disabled in a safe and secure manner. The current concerns are that associated with any new technology and can be overcome with expertise, innovation and of course investment.

2. Externally-driven changes (Access & Inequality) - Various aspects of the SPRITELY design will determine who has access and who doesn't. More funding for ACME robotics would result in better quality designs of SPRITELY. This may lead to greater multimedia applications to give a better level of education and self-learning for the user. The result of this may be more available jobs for disabled through specifically designed learning and development applications, leading to more interest in SPRITELY, leading to eventual price decrease, ultimately resulting in more access.

3. Internally-driven changes (Human-Technology Interaction) - The growth of discretionary robotics, and their adaptability to a wide range of disabilities are more directly tied to the quality of their interfaces than in the past. It would be of far greater benefit to ACME Robotics to develop an interface that could be adapted to any particular disabled persons requirements.

4. Monitoring changes (Control) - An assertive control of overall Project Management should be taken to ensure the final product meets its objective capabilities within the budgetary limitations. HTI control of SPRITELY needs to be improved upon to ensure its safety and reliability during its use by all its potential users.

The group conclusion regarding the SPRITELY article clearly indicated inaccuracies in its published claims. E.g. undergoing final testing with the Health and Safety Executive. ACME will need to develop a comprehensive network of information gathering sources in order to ensure that they are appraised of all the factors that are relevant to manufacturing decisions.

References

• 1. http://www.hse.gov.uk - Health and Safety Executive Web Site visited 7-Sep-2001
  2. http://www.hsl.gov.uk - Health and Safety Executive Laboratory Web Site visited 7-Sep-2001
  3. DECT phone manual - BT Quartet Manual
  4.'The individual at the centre' - Marika Karlsson, Karin Jönsson and Ulf Larsson - CDROM
  5. 'An investigation of the Therac-25 accidents' - Leveson, N. G. and Turner, C. S. (1993)-CDROM.
  6. 'Evaluation of safety-critical software' - David L. Parnas, A. John van Schouwen, Shu Po Kwan - CDROM
  7. 'Software safety: why, what, and how'. - Nancy G.Leveson - CDROM
  8. 'Jobs for Bots' - New Scientist
  9. 'Software safety: why, what, and how'. - Nancy G.Leveson - CDROM
  10.'Jobs for Bots' - New Scientist
  11. 'One Hundred Percent Assured Performance for Robotic Devices for Handicapped and Elderly' - S. E. Blatt; H. B. Mairan - CDROM

Further Information

The article 'Jobs for Bots' from the New Scientist may be found online at: http://www.newscientist.com/hottopics/ai/jobsforthebots.jsp
The robot CYE is being developed by Probotics Inc. in America and more information can be found at http://www.personalrobots.com/home.html