Robomaster

I implemented robotic controlling functions and CV algorithm in RoboMaster team.

RoboMaster is a robotic competition. In short, the rule is to control our robots to win the game, by shooting enemies’ armor (4 black panels fixed at 4 sides of chassis) with specific plastic bullets.

When I joined the newly-formed RoboMaster team, I collaborated with another teammate and developed motion controlling algorithms on our robot at first. We made some advanced motion functions on our robot, such as “Dancing Mode”. In this mode, we fixed the gimbal while make chassis pivot, so that the armor on chassis won’t be shot while we can still shoot enemy. Moreover, the robot itself can still move in dancing mode.

After that, we turned to the advanced part - CV. RoboMaster Competiton encourages participants to explore CV algorithms by introducing a subgame. The upper screen shows password with 5 regular digits while 9 lower screens show unique hand-written digits from MNIST dataset, and both upper and lower screens refresh every 1.5 seconds. During the competition, one that successfully shoots 5 MNIST digits corresponding to the password can win a double attack for the whole team, which is an important factor to win. We utilized OpenCV and tiny_dnn to perform detection, segmentation and recognition. To ensure our robot’s shooting accuracy, we performed camera calibration, PnP solving and physical error correction. Finally, we optimized our program with multithreading and network pruning for onboard computer. All these efforts led to 98% overall accuracy. Our code can be found here.

After 2017’s competition, I designed the CV challenge for new members and wrote the judge program including server and client. Members need to recognize the digit on the screen and send their results to server through client.

I then became the head of CV group, organized several internal meetings for members to report their inspiring ideas, and helped them do research on wider CV applications such as enemy detection and tracking.