Zhaoting Li

CARE: Contact Aware and Reactive Estimator for Contact-Rich Motion Planning 2022.10 - current

Ongoing, MSc thesis in Computational Robotics Lab (CRL), ETHz

• Implemented a contact detection algorithm that can estimate the contact locations and fores using the data from the joint torque sensors.
• Extend the contact particle filter method to output multiple possible solutions, plan to generalize the state of art methods to multiple contact cases.
• Implemement a contact-aware control framework for rigid contacts, and generalize this framework for contacts with unknown stiffness, includling both soft and rigid contacts.
• Plan to propose high-level contact-aware planning in cluttered envrionments to find possible trajectories which may leads to necessary contacts with the environment.

Learning haptic data processing in Human Robot Collaborative Carrying Task 2022.03 - 2022.08

semester project in Robotic System Lab (RSL), ETHz [Report]

• Proposed a learning-based force filtering method that utilizes the filtered-and-amplified robot-side forces as the ground truth, to reduce the noise introduced by the trotting action.
• Designed an efficient transformer-based neural network,collected data during the collaboration between humans and ANYMal with an arm, and trained the network in the typical supervised learning framework.
• Tested the effectiveness of the proposed force filtering method with the admittance control framework.

Learning Muti-Robot Exploration Strategy with Adaptive-QMIX algorithm 2021.10 - 2022.02

based on the robot exploration project

• Developed the simulation environment for muti-robot exploration.
• Proposed the adaptive-QMIX method which can be adaptive to various total robot number.
• Applied the proposed algorithm to muti-robot exploration problem.

Deep Learning for Autonomous Driving: muti-tasking learning and 3D object detection 2022.04-2022.07

Course project of Deep learning for autonomous driving

• Trained neural neworks based on three types of muti-task learning framework, compared the effects of different hyperparameters and structures.
• Proposed an adaptive weight tuning method which can dynamically adjust the weight of each task’s loss during training.
• Trained the baseline proposal refining network, and proposed several methods to improve its performance.

Learning Robot Exploration Strategy with Deep Reinforcement Learning 2020.10-2021.07

Research work at SUSTech, published at RA-L [paper].

• Proposed 4D point-clouds-like information, which consists of 2D points’ location information, and the corresponding 1D frontier and 1D distance information.
• Designed the corresponding training framework based on the deep Q-Learning method and modified it to adapt to the variable action space.
• Demonstrated the performance of the proposed method on a wide variety of environments, which the model has not seen before, and includes maps whose size is much larger than maps in the training set.

Efficient Heuristic Generation for Robot Path Planning 2020.07-2020.10

Research work at SUSTech, published at ICRA 2021 [paper].

• Designed a novel recurrent generative model to generate efficient heuristic for robot path planning.
• Incrementally constructed the heuristic through the feedback of historical information.
• Combinedthe generated heuristic with RRT* algorithm to guide the algorithm to find both initial and optimal solutions in a faster and more efficient way.

A sampling-based motion planning method for urban autonomous vehicles 2019.07-2020.01

Research work at UC Berkeley, published at IFAC 2020 [paper].

• Applied the discrete elastic-band-based motion planning method(EB planner)to generate piecewise linear collision-free path with dynamic programming.
• Employed pure pursuit controller to smooth this path.•Used a spatial and speed sampling methodtogether with acascaded ranking method to optimize the trajectory with many hierarchical features.
• Applied a non-conservatively defensive strategy to avoid overreacting to threats with low probability.
• Adaptively adjust sampling resolution based on the environment and the objectives of the ego vehicle.

A trajectory planning method for robot scanning sytem for scanning objects with unknown model 2019.02-2019.06

Research work at HIT, published at Neurocomputing 2020 [paper].

• Designed an automatic and low-cost robot scanning system consisting of a kinect camera, a UR 10 robot and a line laser scanner. Also calibrated the coordinate systems.
• Proposed an online correction methods based on follow-up control and scanned data to optimize the pose of the laser scanner.
• Proposed a novel path planning methods for laser scanning based on the least square fitting and online correction. This path planning method has been validated in many use cases under various work conditions.