Monopod Jumping Robot

Objective:

Design a jumping robot to compete in a high and long-jump competition, as well as an obstacle course, that meets motor, processing, manufacturing, time, weight, and cost constraints.

Design:

It was decided to design a monopod robot that would jump by pulling a cylinder containing a spring over a piston, pointing the cylinder in the desired direction, and releasing the cylinder. Please refer to this diagram (100 Kb PDF) for an explanation of the mechanism. You can view the detailed strategy here (46 Kb PDF).

A limit on the number and type of motors used was given, so efficiency in use of motor power output was critical to performance. A 90% efficient ball screw was used to compress the spring, instead of a 30% efficient ACME lead screw. Other compression mechanisms could not provide enough compression force and stay within the size constraints. Series and planetary gear reduction was used to increase torque to the ball screw, increasing the available compression force to the range of 200 to 300 pounds (depending upon gear selection). A coupler with thrust bearings was required to allow rotation of the ball screw while the mechanism was latched. You can download the full size robot descriptive photos below:

To reduce weight and stay within size constraints, only one ball screw could be used. To prevent binding that might result from this offset force, a precision linear motion mechanism was developed that incorparated machining tolerances within 0.0005", and precision bearings rolling through grooves in the piston shaft. To release the spring, a latch mechanism was developed similar to a bicycle linear-pull brake, driven by a cable and linear actuator custom made to fit one of the specified motors.

To stay within weight constraints, and to improve performance, many components were custom fabricated out of carbon fiber, Kevlar, or hybrid composites.

To control horizontal direction, a cable driven by a motor rotated the piston on its base to point in the desired direction of travel. The piston shaft was attached to the base via a custom machined deep-groove thrust bearing, as no off-the-shelf bearing to meet the requirements could be found.

The control scheme developed was simplistic, requiring only an encoder on the ball screw, and an array of IR sensors on the top of the robot. A diagram of the control scheme may be found here (88 Kb PDF).

Performance:

Nearly every aspect of the robot worked well, after varying degrees of refinement. The adjustment mechanisms were reliable, the linear motion mechanism never bound, and soon the robot was jumping successively at over one foot of height. However, right before the competition two mishaps occured. One of the carbon-fiber struts broke loose from the base, rendering the base flexible. In addition, after a particularly high jump (18"+), the robot landed in an undesirable position, and broke the encoder. This made it impossible to reset the jumping mechanism, i.e. the robot could jump only once autonomously. This occured just hours before the competition, and not enough time remained to make repairs. In the competition the robot could only make one very large jump in each event. It scored no points in the obstacle course because it could not jump the full 2 meters. However, because it could jump more than 0.5 horizontal meters, it did score points in the other events.

Conclusion:

Not only was this an excellent learning tool in mechanical design and control, but participants learned a great deal about the importance of effective project management, and the woes of scope creep. Had the scope been smaller, or had more resources been available, more testing time would have been available to improve the reliability of the robot.