Boadicea - Conclusions

Many researchers have built legged robots. In general, however, these machines have lacked the mechanical capabilities required to present an advantage over wheeled and tracked machines. The robots are orders of magnitude slower than wheeled vehicles, and have limited payload capacities. These shortcomings must be addressed before legged locomotion will be a viable alternative.

Many robot designs have been severely handicapped by their actuators. Electric motors are impractical for mobile robot applications; their torque densities are too low. Boadicea has indicated the performance advantages pneumatic actuators provide for small legged robots. Although fluid power systems are less familiar to robot designers,, the performance improvements are worth the effort required. Future legged robots should use fluid power actuators to improve their mobility.

Legged machines in the future should also make use of force and impedance control modes. These control modes could simplify controlling body velocity and orientation; forces exerted by the legs directly determine body motion. Practical dynamic locomotion strategies will likely require force or impedance control. Although Boadicea was unable to explore these possibilities, its pneumatic actuation system would make them easier to implement than with an electric system.

Insects and other multi-legged animals do not operate under the assumptions normally made by robot designers, including maintaining static stability, moving the body horizontally at constant velocity, and making each leg identical. Studies of animal locomotion reveal advantages that can be achieved by dispensing with these constraints [Full, et al 89]. Boadicea demonstrates some of these improvements.

Experiments with dynamic robots and results from studies of insect locomotion [Full and Tu 91] illustrate the speed advantages of dynamic locomotion. Energy efficiency can also be improved with dynamic gaits; storing allows it to be recovered on the next stride [Cavagna, et al 77]. Generating dynamic motion in a small robot like Boadicea would require more detailed modeling of leg, body, and actuator dynamics. Higher performance sensors would also be required. To date, results from dynamic robot experiments have been encouraging, and future legged robots should continue in this direction.

Another difference between Boadicea and its predecessors is that Boadicea does not have six identical legs. I attempted to incorporate some of the optimizations found on insect legs to improve the robot's capabilities. Boadicea's leg designs offer some mobility advantages, although not as significantly as I expected. Climbing experiments with the robot indicate that the rear and middle legs are not specialized enough. The middle legs need to lift the body more, and the rear legs should provide more propulsion.

This page is still under construction.

[Boadicea top page]
[Fluid Power] [Materials & Construction] [Biological Inspirations]
[Leg Design] [Sensors & Electronics] [Control System]
[Results] [Conclusions]
[Photos & Video] [Papers] [Related Web sites]