RESEARCH

SEISMIC SENSOR DEPLOYMENT SIMULATION When it came time to animate the TETwalker, the newly designed model would not work properly. The linear actuators would lengthen to extreme distances and push different parts of the model off the ground. This is an inherent challenge and difficulty with the robot architecture and design in that for one strut to move, multiple other struts must also move. The second model of the TETwalker was created using rigid cylinders rather than using linear actuators for the base design for the tetrahedron. We did not want the base of the TETwalker to push in and out but rather stand still to demonstrate how the robot topples for mobility. The next objective was to simulate the actual deployment of the seismic sensor. Also, spheres replaced the usage of cubes as the nodes. The TETwalker could maneuver easier while using spheres rather than cubes since they would roll rather than move piece by piece as the TETwalker would topple over. Unlike the previous model, an object was needed in order to hold all the parts together. Shorter cylinders were created and the struts and nodes were connected with revolute joints. All spheres and cylinders were arranged in the same areas as nodes and struts in the previous model. By the time all the parts were in place, it was time to figure out how to get the center strut to plant and retrieve the seismic sensor. Alteration of the linear actuator properties was necessary in order to approximate the time in which the actuator would need to move up and down. Timing of strut movement was critical in the design as the need to extend/retract a single strut requires many other struts to also change in length or pivot. A table of lengths and times to achieve those lengths were created to control the behavior of each strut using a cubic spline interpolation function. This controlled how fast each strut extended and retracted to perform the toppling motion. The final product of the actuators resulted in the times of 0, 0.5, 1.0, 1.5, 2.5, 3.0, 3.5, 4.5, and 5.0 seconds. The actuators would correspondingly move to lengths of 0.1, 0.035, 0.1, 0.1, 0.035, and 0.075 meters at those times. Each actuator rises for a short period of time before it then lowers itself to the ground, simulating the robot’s motion of actually deploying the sensor onto/into the surface. The actuator continues to rise to retrieve the sensor, and the above steps are repeated for each deployment location. Fig. 5 shows images of the deployment and retrieval simulation. Click Here to See Seismic Sensor Deployment Simulation Movie Clip MOBILITY SIMULATION MSC.visualNastran was used to simulate mobility as it integrates lifelike physics. The TETwalker travels by moving its center node to one side until the unbalanced weight topples in the direction due to the gravity feature in the program. Originally, we removed all of the center pieces of the TETwalker and made all the struts into linear actuators. We attempted to lengthen two of the three struts connected to the center node. This resulted in the two struts pushing the node over the shortened one. In order to shorten and lengthen all actuators, we created a table for each in the software which consisted of the distance in which the strut would lengthen or shorten and the amount of time to perform the action. For the two nodes that would be lengthening, we set the lengths as 0.288, 0.7, and .07 meters and respectively set their times to 0, 1, and 3 seconds. For the shortened strut, we set the lengths as 0.277, 0.377, and 0.377 meters at the times of 0, 1, and 3 seconds. After the two struts push over the shortened one, the newly placed struts in the middle push up together to raise the center node again. Two out of the three nodes would lengthen to 0.361, 0.261, 0.361, 0.561, and 0.561 meters at the times of 0, 0.75, 1.25, 1.5, and 2.25 seconds while the other one stayed the same length. When viewing the simulation, the robot’s movement was very choppy and slow since it was having difficulty raising all the nodes on its own. The linear actuators must be in a stable enough position so that the robot can bring itself back into the upright and stable position to insure the center of gravity can properly be shifted using the available linear actuator lengths. Fig. 6 shows the extension of the top node to alter the center of gravity in order to topple. We decided to focus more on modeling a single toppling motion to better understand system dynamics. This design consisted of three linear actuators, just enough to extend a node to alter the center of gravity. Out of the three linear actuators, only one actuator was required to lengthen while the others to remain the same length. This actuator was set on the times of 0, 0.25, and 0.36 seconds that correspond with the lengths of 0.173, 0.25, and 0.36 meters. To assist in the TETwalker’s toppling motion, the top node’s weight had to be altered in order for the transition to work effectively. Originally, the weight of the center node was set at 0.1 kilograms, but needed to be changed to 4 kilograms to successfully topple. When the simulation was finalized, the actuator would extend itself enough to actually turn the TETwalker over while pulling another node in its direction, allowing the robot to continue to topple. Click Here to See Mobility Simulation Movie Clip