ASTOR V1 [Image from "Cybernetics" , John Frederick Young, 1969]
As from the text below, ASTOR V1 used discrete components and was built by P. A. Kidd and C. J. Lloyd. It was probably built earlier than the book date, but I have not found any other record of this.
ASTOR V2 [ no image] used integrated circuits and was built by B. J. Alston and I. Foxall [no date].
-x640.jpg)
Text from "Robotiics", John Frederick Young, 1973
Simple mobile machines pp18-22
In order to be able to investigate problems or reliability in mobile robots at Aston [University, Birmingham, England], the performance of small 'tortoise' machines is being investigated. Early models of such machines had a poor reliability, and it has been possible to improve this by a study of the reasons for each failure.
The main points of interest in such a study are:
1. The operation of electronic control and logic circuits on a battery supply.
2. The necessary size of batteries and the required frequency of recharging and rate of recharging.
3. The interaction between different parts of the control circuitry when operated on common supplies.
4. The possibility of arranging the robot to plug itself into a charging supply whenever necessary. (This can fancifully be thought of as `feeding'.)
5. The effect of a mobile environment on the control equipment and on the moving parts of the device.
6. Problems of cleanliness in a mobile environment.
7. The required speed of movement.
8. Methods of conveying to the mobile robot information about its environment.
9. Efficient methods of control of the motive power.
10. Safety factors such as obstacle-detecting switches, emergency stops and their locations on the robot.
11. Rapid and efficient methods of dismantling for maintenance which at the same time are 'child-proof'.
12. Possibilities of redundancy of parts and the acceptance of reduced performance in emergencies. Is it preferable to stop in the event of any failure or to continue operation despite a reduced performance?
13. Possibilities of self-repair.
14. Noise produced by the robot.
Robot Stability p103
In order to be able to investigate problems or reliability in mobile robots at Aston [University, Birmingham, England], the performance of small 'tortoise' machines is being investigated. Early models of such machines had a poor reliability, and it has been possible to improve this by a study of the reasons for each failure.
The main points of interest in such a study are:
1. The operation of electronic control and logic circuits on a battery supply.
2. The necessary size of batteries and the required frequency of recharging and rate of recharging.
3. The interaction between different parts of the control circuitry when operated on common supplies.
4. The possibility of arranging the robot to plug itself into a charging supply whenever necessary. (This can fancifully be thought of as `feeding'.)
5. The effect of a mobile environment on the control equipment and on the moving parts of the device.
6. Problems of cleanliness in a mobile environment.
7. The required speed of movement.
8. Methods of conveying to the mobile robot information about its environment.
9. Efficient methods of control of the motive power.
10. Safety factors such as obstacle-detecting switches, emergency stops and their locations on the robot.
11. Rapid and efficient methods of dismantling for maintenance which at the same time are 'child-proof'.
12. Possibilities of redundancy of parts and the acceptance of reduced performance in emergencies. Is it preferable to stop in the event of any failure or to continue operation despite a reduced performance?
13. Possibilities of self-repair.
14. Noise produced by the robot.
Robot Stability p103
It can be seen from this list, which is not at all a complete one, that study of small mobile robots can help to clear up some of the problems even before they are encountered on full-size, mobile, working robots. Some of the problems can in fact be studied on non-mobile equipment, but a final study on mobile equipment is desirable.
The simple mobile machine Astor
Several experimenters have used the form of the beetle or of the tortoise for simple mobile robot machines. Perhaps one of the best publicised has been that due to Walter18. This form was later elaborated by Angyan19.
In the Aston Cybernetics Laboratory experiments have been carried out with the small machine Astor, which also has the general shape of a tortoise22. The mechanical arrangement was first produced by P. A. Kidd and C. J. Lloyd, who adopted the use of transistors in the control system in an attempt to improve the reliability of the earlier forms.
The early model has three wheels, two at the rear and one centrally situated at the front. The rear wheels are free-running. The front wheel has two operating motors. One of these, the drive motor, causes the wheel to drive the model along. The second, the scan motor, causes the wheel to rotate about its vertical axis, so giving a steering action. All of the possible actions of the device are caused by the action of these two motors.
The movement of the device is influenced by inputs from two photocell 'eyes' at the front, by a microphone 'ear', and by a movable shell which can detect the presence of obstacles by touch.
The model is designated to operate as follows :
1. If there is no external stimulus, the steering operates in such a way that the device searches for a light source.
2. If any obstacle is encountered, the model backs away before resuming its search for a light source.
3. If a weak light is detected by one of the photocells at the front of the device, then the motors operate to drive in the direction of the light.
4. If a strong light is detected by the other photocell at the front of the device, then the motors operate to drive away from the light.
5. If the sound of a whistle is detected by the microphone, then the model stops completely, or 'freezes' for a time.
6. If a certain number of coincidences occur of weak light detection and the detection of a whistle sound, then conditioning occurs and for a time the sounding of a whistle has exactly the same effect as the detection of a weak light source.
In order to achieve these modes of action, the Astor device is connected as follows :
1. To simulate a searching mode, both motors are energised simultaneously. The motor speeds are regulated to ensure that, although the model moves about in an apparently random fashion, a complete 360° arc of search is covered in time.
2. To avoid an obstacle, the drive motor is switched off for a short period whenever an obstacle is encountered. This gives the scan motor time to turn the wheel away from the direction of the obstacle before drive recommences.
3. Whenever the photocell at the front of the device detects a weak source of light, the drive motor is immediately switched off but the scan motor continues to operate until the drive wheel is aligned in the direction of the light. The scan is then switched off and the drive motor is reconnected so that the device moves in the direction of the light source. There is thus a waiting period while the steering is aligned.
4. If the light detected by the photocells becomes too strong, the scan motor is reconnected and the search for a moderate source of light restarts.
5. Both motors are switched off if a whistle is sounded, so causing the model to 'freeze'.
6. In order to simulate a conditioned reflex action, the weak light input and the sound input are taken to a coincidence detector, the output of which goes to a counter circuit. The counter gives an output only after the weak light and the whistling sound have coincided a number of times. The output then gradually fades, if it is not reinforced, to simulate the forgetting feature. While the coincidence memory output is present, any further soundings of the whistle are temporarily arranged to act as though an input from a weak light source had been received.
The over-all block diagram of this control system is shown in Figure 6.7. This is composed mainly of integrated circuits and was designed and constructed by B. J. Alston and I. Foxall of the Cybernetics Laboratory at Aston, to replace an earlier scheme using discrete components22.
The work on these schemes has revealed the sort of problems which are likely to be found with mobile robot devices. Later work has been based on this early work, and is aimed at the mounting of
104 Robot Stability
the Astra mark 3 machine on a mobile trolley, in order to investigate further the problems involved with the mobile and independent robot.
REFERENCES
1. Young, John F., 'Simplified Feedback Stabilisation', Process Control Automn, 10, June, 233 (1963).
2. Young, John F., 'The Roots of Quadratic Equations', Control, 7, November, 237 (1963).
3. Young, John F., 'Electronic and Magnetic Amplifier Voltage and Frequency Regulators', Electricity in Industry, No. 9 (1956).
4. Mazda, F. F., 'An Electric Vehicle Controller', Electron. Components, 12, March 19, 235 (1971).
5. Andriesse, R. D., 'Split Field Servo Motors', Control, 9, August, 425 (1965).
6. Anon., 'Characteristics of British Servo Motors', Automn Prog., June, 198 (1960).
7. Williams, F. C., 'The Velodyne', in: Servomechanisms, H.M.S.O., 134 (1951).
8. Williams, F. C. and Uttley, Albert M., 'The Velodyne', JIEE, 93, pt 3A, 1256 (1946).
9. Lampert, W. E. C., 'Naval Applications of Electrical Remote Position Controllers', JIEE, 94, pt 2A, 236 (1947).
10. Taylor, P. L., Servomechanisms, Longman, 193 (1960).
11. West, J. C., Servomechanisms, English Universities Press, 182 (1955).
12. Young, John F., 'The Boucherot Effect', Wireless Wld, 68, August, 391 (1962).
13. Young, John F., 'Voltage Regulator Comparison Circuits', Control, 6, June, 90 (1963).
14. Young, John F., 'Field Control of Motors with Constant Current Armature Supply', Control, 12, January, 35 (1968).
15. Young, John F., Ferroresonance; Problems and Aplications', Elect! Rev., 176, May 21, 782 (1965).
16. Young, John F., Bibliography on Ferroresonance, deposited with Library at University of Aston.
17. Young, John F., 'Using Capacitors to Improve Low-Cost Magnetic Amplifiers', Engineer, Lond., 220, July 30, 176 (1965).
18. Walter, W. G., The Living Brain, Duckworth (1953).
19. Angyan, A. J., 'Machina Reproducatrix', in: Mechanism of Thought Processes, Vol. 2, 933, H.M.S.O. (1959).
20. Zemanek, H., et al., 'A Model for Neurophysiological Functions', in: Cherry, C. (ed.), Information Theory, 1960, 270, Butterworths (1961).
21. Kukhtenko, A. I., 'The Dynamics of Devices Which Imitate Living Organisms', in: Coales, J. F. (ed.), Automatic and Remote Control, Vol. 2, 658, Butterworths (1961).
22. Young, John F., Cybernetics, Wife (1969).
23. Young, John F., 'Control Possibilities of Double-Stator Squirrel-Cage Motors, Control, 12, May, 416 (1968).
24. Young, John F., 'Phase Measurements in Feedback Amplifiers', Electron. Eng., 27, July, 311 (1955).
25. Nightingale, J. M. and Todd, R. W., 'Adaptive Control of a Multi-Degree of Freedom Hand Prosthesis', Proc. Conf. Hum. Locomotor Eng., Sussex, September, 1971 (Instn Mech. Engrs), 249.
26. Nightingale, J. M., 'Intelligent Prostheses', Proc. Meeting on Uses of
Robots, London, April, 1973, 3
The simple mobile machine Astor
Several experimenters have used the form of the beetle or of the tortoise for simple mobile robot machines. Perhaps one of the best publicised has been that due to Walter18. This form was later elaborated by Angyan19.
In the Aston Cybernetics Laboratory experiments have been carried out with the small machine Astor, which also has the general shape of a tortoise22. The mechanical arrangement was first produced by P. A. Kidd and C. J. Lloyd, who adopted the use of transistors in the control system in an attempt to improve the reliability of the earlier forms.
The early model has three wheels, two at the rear and one centrally situated at the front. The rear wheels are free-running. The front wheel has two operating motors. One of these, the drive motor, causes the wheel to drive the model along. The second, the scan motor, causes the wheel to rotate about its vertical axis, so giving a steering action. All of the possible actions of the device are caused by the action of these two motors.
The movement of the device is influenced by inputs from two photocell 'eyes' at the front, by a microphone 'ear', and by a movable shell which can detect the presence of obstacles by touch.
The model is designated to operate as follows :
1. If there is no external stimulus, the steering operates in such a way that the device searches for a light source.
2. If any obstacle is encountered, the model backs away before resuming its search for a light source.
3. If a weak light is detected by one of the photocells at the front of the device, then the motors operate to drive in the direction of the light.
4. If a strong light is detected by the other photocell at the front of the device, then the motors operate to drive away from the light.
5. If the sound of a whistle is detected by the microphone, then the model stops completely, or 'freezes' for a time.
6. If a certain number of coincidences occur of weak light detection and the detection of a whistle sound, then conditioning occurs and for a time the sounding of a whistle has exactly the same effect as the detection of a weak light source.
In order to achieve these modes of action, the Astor device is connected as follows :
1. To simulate a searching mode, both motors are energised simultaneously. The motor speeds are regulated to ensure that, although the model moves about in an apparently random fashion, a complete 360° arc of search is covered in time.
2. To avoid an obstacle, the drive motor is switched off for a short period whenever an obstacle is encountered. This gives the scan motor time to turn the wheel away from the direction of the obstacle before drive recommences.
3. Whenever the photocell at the front of the device detects a weak source of light, the drive motor is immediately switched off but the scan motor continues to operate until the drive wheel is aligned in the direction of the light. The scan is then switched off and the drive motor is reconnected so that the device moves in the direction of the light source. There is thus a waiting period while the steering is aligned.
4. If the light detected by the photocells becomes too strong, the scan motor is reconnected and the search for a moderate source of light restarts.
5. Both motors are switched off if a whistle is sounded, so causing the model to 'freeze'.
6. In order to simulate a conditioned reflex action, the weak light input and the sound input are taken to a coincidence detector, the output of which goes to a counter circuit. The counter gives an output only after the weak light and the whistling sound have coincided a number of times. The output then gradually fades, if it is not reinforced, to simulate the forgetting feature. While the coincidence memory output is present, any further soundings of the whistle are temporarily arranged to act as though an input from a weak light source had been received.
The over-all block diagram of this control system is shown in Figure 6.7. This is composed mainly of integrated circuits and was designed and constructed by B. J. Alston and I. Foxall of the Cybernetics Laboratory at Aston, to replace an earlier scheme using discrete components22.
The work on these schemes has revealed the sort of problems which are likely to be found with mobile robot devices. Later work has been based on this early work, and is aimed at the mounting of
104 Robot Stability
the Astra mark 3 machine on a mobile trolley, in order to investigate further the problems involved with the mobile and independent robot.
REFERENCES
1. Young, John F., 'Simplified Feedback Stabilisation', Process Control Automn, 10, June, 233 (1963).
2. Young, John F., 'The Roots of Quadratic Equations', Control, 7, November, 237 (1963).
3. Young, John F., 'Electronic and Magnetic Amplifier Voltage and Frequency Regulators', Electricity in Industry, No. 9 (1956).
4. Mazda, F. F., 'An Electric Vehicle Controller', Electron. Components, 12, March 19, 235 (1971).
5. Andriesse, R. D., 'Split Field Servo Motors', Control, 9, August, 425 (1965).
6. Anon., 'Characteristics of British Servo Motors', Automn Prog., June, 198 (1960).
7. Williams, F. C., 'The Velodyne', in: Servomechanisms, H.M.S.O., 134 (1951).
8. Williams, F. C. and Uttley, Albert M., 'The Velodyne', JIEE, 93, pt 3A, 1256 (1946).
9. Lampert, W. E. C., 'Naval Applications of Electrical Remote Position Controllers', JIEE, 94, pt 2A, 236 (1947).
10. Taylor, P. L., Servomechanisms, Longman, 193 (1960).
11. West, J. C., Servomechanisms, English Universities Press, 182 (1955).
12. Young, John F., 'The Boucherot Effect', Wireless Wld, 68, August, 391 (1962).
13. Young, John F., 'Voltage Regulator Comparison Circuits', Control, 6, June, 90 (1963).
14. Young, John F., 'Field Control of Motors with Constant Current Armature Supply', Control, 12, January, 35 (1968).
15. Young, John F., Ferroresonance; Problems and Aplications', Elect! Rev., 176, May 21, 782 (1965).
16. Young, John F., Bibliography on Ferroresonance, deposited with Library at University of Aston.
17. Young, John F., 'Using Capacitors to Improve Low-Cost Magnetic Amplifiers', Engineer, Lond., 220, July 30, 176 (1965).
18. Walter, W. G., The Living Brain, Duckworth (1953).
19. Angyan, A. J., 'Machina Reproducatrix', in: Mechanism of Thought Processes, Vol. 2, 933, H.M.S.O. (1959).
20. Zemanek, H., et al., 'A Model for Neurophysiological Functions', in: Cherry, C. (ed.), Information Theory, 1960, 270, Butterworths (1961).
21. Kukhtenko, A. I., 'The Dynamics of Devices Which Imitate Living Organisms', in: Coales, J. F. (ed.), Automatic and Remote Control, Vol. 2, 658, Butterworths (1961).
22. Young, John F., Cybernetics, Wife (1969).
23. Young, John F., 'Control Possibilities of Double-Stator Squirrel-Cage Motors, Control, 12, May, 416 (1968).
24. Young, John F., 'Phase Measurements in Feedback Amplifiers', Electron. Eng., 27, July, 311 (1955).
25. Nightingale, J. M. and Todd, R. W., 'Adaptive Control of a Multi-Degree of Freedom Hand Prosthesis', Proc. Conf. Hum. Locomotor Eng., Sussex, September, 1971 (Instn Mech. Engrs), 249.
26. Nightingale, J. M., 'Intelligent Prostheses', Proc. Meeting on Uses of
Robots, London, April, 1973, 3
There was a later model, referred to as ASTRA Model 3 that was based on artificial nerve cells.
2 Replies to “1969 – “Astor” – Aston Cybernetics Laboratory (British)”
Comments are closed.