Journals: Adam_Carol

Week 1:


This marks the start of our project. We evaluated our starting point and set up a meeting with Dr. Koditschek


Today, we spent class time meeting with Dr. Koditschek to discuss the feasibility of our project and potential project end goals. Before speaking with Dr. Koditschek, we had only a general project scope in, it became clear that we needed a much more refined hypothesis to test and an 'null' baseline hypothesis, which represents the default assumption. We shifted our focus on creating a five legged gait that performs better than the five legged tripod gait, with the default assumption that the five legged tripod gait performs badly. With a better formulated experiment, we adjusted our plans for the upcoming weeks. We would first observe Junior performing the five legged tripod gait.

Week 2:


We performed additional literature research to give us more ideas on how to change the five legged tripod gait. We were able to identify several key adaptations in leg gait that animals show after they sustain an injury to leg locomotion. Variations in leg sweep and stance angles and load bearing were observed across different animals (rats and cockroaches) after leg injuries and amputations. Relative movement in a leg adjacent to the injured leg versus a leg on the other side also varied. This suggests that we need to alter Buehler clock parameters individually for legs. Researchers also looked at specific metrics in evaluating gaits - acceleration, roll, pitch.


We identified a lab with the Vicon system. After preliminary research, we are prepared to implement it, anticipate the nature of the collected data, and analyze the data. We also set up a meeting to learn how to setup the robot.

Week 3:


- Learned how to setup Junior platform from Deniz
- Continued coding in preparation of data collection later this week.



We met and revised our plans for the remainder of the semester. Specifically, we narrowed down the data we planned to collect to evaluate the performance of each gait, discussed a systematic procedure for developing our new gait and collecting the appropriate data, and decided on the appropriate formats for the presentation of our findings. We also revised our time line for the completion of each of our remaining tasks. This included meetings on Saturday and the following Tuesday for gait implementation and data collection as well as a time line for reading the literature to gain a better conceptual understanding and to improve our experimental methods.


We met to run experiments and collect data from the robot. First we wrote the code to read and store the desired imu and current data while running the robot along a straight path. We also figured out how to export arrays and save them in .mat files to allow us to process the data in MATLAB. We verified that we were able to generate time histories of yaw, pitch, roll, and current and store them as MATLAB arrays. We ran trials where we rotated the robot about each axis individually in order to ensure that the data we were saving was reflective of the robot's motion. We then ran trials of a six legged tripod gait as well as the same gait with leg 1 stationary in flight in order to collect data. We will be comparing the results from our new five legged solution against this data. We plan to use the motor current data to determine when each leg is on the ground for each of our trials.


We read through more of the literature in order to improve our approach to developing our five legged gait. We particularly looked at cockroach responses to leg amputations. We did some Python coding to prepare for the next round of experiments and some MATLAB coding to prepare for data processing.