8.0 CENTRE OF MASS ROBOT
During the study of walking robots a completely different method of walking locomotion was derived. Chapter 2 mentioned how very few robot designers ever moved the centre of mass of their robots. This was opposed to the findings in chapter 3 where almost all animals even insects moved their centre of mass. It became apparent that to move the centre of mass, allot, could be very advantageous.
By using a moving center of mass, a completely different SCR is developed.
The reasoning was based on making the simplest possible walking device. The design is shown in figure 8.1 Joints A, B, C & D are actuated.
Figure 8.1 lay out of COMR (Centre of Mass Robot)
Figure 8.1 continued
The method of locomotion is as follows.
1) move mass to one end such that the sum of moments around pivot A is 0 or slightly -ve around A.
2) Cause actuator at joint A to rotate to desired position (this raises foot 1)
3) rotate joint B to through desired angle
4) remove actuation force at joint A
5) drive counter weight to other end
6 )repeat this process with the other foot (foot Q)
A demonstration of this is supplied in the second segment of the Video.
This process allows the robot to climb to two surfaces within the range defined by the angle of incline and length between the two feet.
A prototype was built out of Technical Lego and tested.
This approach has some major advantages over conventional walking robots.
1 There is no complex leg placement.
2 walking around corners is very easily achieve as the robots next step can be in any direction, this removes all the complicated leg Co-ordination, for turning corners or through any angle.
3 The vertical height that the robot can achieve is quite remarkable when compared with other small and even large robots, it easily out performs any comparably conventionally sized robot .
4 Almost any surface could be accommodated by allowing three screw type feet on each foot.
5 The power supply was part of the pay load, it was made into a useful component instead of a liability, as in most robots. It serves as part of the counter weight mass.
6 Climbing is very efficient, their re no complex positions that the body has to assume to keep balance.
7 There is lots of scope for development. This model is only a prototype. The practicality of the idea is demonstrated in the time it took to implement, which was about 3 hours.
8 Control systems and strategies, would be much simplified, due to the reduction in d.o.f and the way in which this type of robot navigates terrain.
This robot design achieves walking, and the tested robot can almost climb a step.
Its limitations stems from the physical layout of the robot. That is the counter weight, when at one end limits how close the robot may move towards a wall. A dynamic running motion would be hard to achieve., but again wheels could be used on the feet to drive over smooth terrain.
This entire idea is quite a radical departure from conventional robot designs, however it demonstrates an alternative method of type 3 terrain navigation. Several advances could be made to improve the robots performance.
1) Use a heavier counter weight.
2) Redesign the robot so that is can rotate to the position shown in figure 8.03.
3) Make the connecting beam element 3 telescopic, this would reduce the distance the counter weight has to travel climb small distances, thus making the robot more efficient.
4) Make the robot lighter, use aluminium and carbon fibre elements instead of Lego.
5) Redesign the robot such that one foot may be place vertically over the top of the other foot.
By combining improvements 3 ,4 and 5 the counter weight and thus the total power consumption of the robot may be greatly reduced, as well as improving the maximum height that the robot can climb. It would be quite feasible to make a robot of similar proportions to the six legged robot able to make a vertical step of over 1 m. See figure 8.2 to see how the turning circle can be reduced to the size of the base foot. Figure 8.3 shows how the improvements would argument the COMR's vertical performance.
Figure 8.2 .This shows how the turning circle may be reduced to the size of the base foot
Figure 8.3 shows how the improvements could argument the climbing ability. (1) The feet are place vertically in line and the telescopic tubing is also vertically aligned, this is to avoid buckling upon extension. (2) & (3) The telescopic tubing is extended (4) The upper foot is place. (5) the counter weight is moved to the upper platform. (6) & (7) the lower foot and tubing is retracted to assume the position in (1) again.
The pay load of the robot would be accommodated as part of the counter weight and feet. To compensate for a change in payload the counter weight would vary the distance it travels beyond the pivots.
The success of this method warrants further investigation of this approach to a walking robot. It yields a solution that circumvents many of the traditional problems that robot designers have faced, such a gaits and leg design.
Copyright, this page is © 1999 TOM EDWARDS, XTCE.COM(excepting TOPOTRICS)