Development of Biomimetic Robotic Fish

This research aims to design a biomimetic robotic fish with hydrostatic stability structure by referring to the motions of real fishes in underwater. Both of the SLA material and EPDM rubber technology are used to achieve the excellent waterproof capability of robotic fish, hence prolong the lifespan of servo motors installed at the robotic fish’s joints. A turning motion model is also designed for the biomimetic multi-joint robotic fish, where nine motion parameters are trained using artificial neural network. The turning angel of each joint is then obtained using geometric optimization and cubic Hermite interpolation techniques. Experiments show that the developed biomimetic robotic fish is able to swim naturally like a real fish using the optimized turning motion model.

Development of Autonomous Vehicle

This research aims to design autonomous vehicle that play crucial roles in the smart transportation systems of city. It is envisioned that each user is able to summon the autonomous vehicle through the service system developed in Web/App. The nearest autonomous vehicle is instructed by the central control system to serve the user’s request. Numerous sophisticated technologies are used to develop this autonomous vehicle, including computer vision and machine learning. Particularly, a Light Detection and Ranging (LiDAR) system is used to create the real-time 3D point cloud of surrounding environment and a stereo depth camera is used to perform image recognition. Finally, the graphics processing unit (GPU) with CUDA platform is also used to facilitate the 3D image processing that is computationally expensive.