About Projects CV Unforgettable-Moments 🌙
Peiyu Yang (杨沛禹)

Welcome to projects!

The links to each project are still under construction, so feel free to contact me for a verbal presentation!

Research

Project 1

Safe Locomotion for Quadrupedal robots

  • We propose a coupled terrain, state, and contact estimation system using proprioceptive sensing.
  • Provide a CBF-based safe motion guarantee relying only on proprioception.
  • I'm currently exploring reinforcement learning and model-based approaches to safety assurance.
  • Thesis project at TU Delft.
Project 1

Design of XingT: A Human-sized Heavy-duty Bipedal Robot

  • Designed a loadable and multi-mode bipedal robot with a height of range 0.9-1.2m.
  • We built the robot hardware and performed extensive simulations and hardware experiments on the subsystems and the whole.
  • The remarkable performance of the robot is demonstrated by the strong load capacity and the active and passive flexibility of the structure.
  • Interest research(on design) and thesis(on EKF) project at BIT.
Project 1

Design and Fabrication of a Legged Robot Prototype: Phase II

  • Designed a self-adaptive robotic leg based on Hoecken’s linkage and pantograph.
  • Proposed a damping compliant method for adaptive structure of the pantograph, which can significantly improve the flexibility of the adaptive structure.
  • Internship project at École Polytechnique de Montréal
 Introduction Slides
Project 1

Target Detection and Kinematic Reconstruction Based on Computer Vision

  • Designed a feature recognition and motion reconstruction system based on orthogonal vision.
  • Education project at North Carolina State University

CourseWork

Project 1

Two-stage Multi-UAVs Planning and Control

  • Presented a two-stage multi-UAV path planning solution with modified A* as global planner, and MPC combining with APF as local planner.
  • Simulated formations of a cluster comprising 27 UAVs in an obstacle environment.
  • Course project at PDM and MPC at TU Delft
 Video Code Planner PDF MPC PDF
Project 1

Intelligent Control Methods for Robotic Arms

  • Train LNNs (Lagrangian Neural Networks) for obtaining the dynamic model of a 2-DoF robotic arm.
  • Apply PD and PD+ methods to the learned model.
  • Course project at ICS at TU Delft.
Project 1

Intelligent Vehicle Navigation Method Based on Evolutionary Neural Network

  • Proposed an evolutionary neural networks (ENN) for vehicle navigation system based on LIDAR.
  • Carried out the simulations, and results showed that the vehicle model can achieve full path tracking operation in about 10 generations in the regular track, and can achieve full path tracking in about 50 generations in the sinusoidal track.
  • Course project at IIC at BIT.

OtherWork