Our research include: Environment-robust uncalibrated visual servoing; Visual contour servoing aiming at object assembly and monitoring; Vision-based automatic manipulation of soft tissue aiming at robot surgery; Vision-based quadrotor control.

Uncalibrated Visual Servoing

An uncalibrated depth-independent adaptive visual servoing system was proposed, which has strong robustness to unknown environment.

Representative work: Uncalibrated vision-based formation control of multi-robots: In a complex environment, select a visual feature of the leader robot to control the motion of follower robots.

• Visual Contour Servoing

For object without obvious visual feature, select its contour feature as the visual feedback of the visual servoing system to control the manipulator. Application field: grasping the object, parts inspection, cutting, assembly, adjustment of monitoring view angle etc.

• Vision-based automatic manipulation of soft tissue

Aiming at surgery robot, research involves deformation and cutting control of soft tissue. Vision-based method is studied to improve the control system’s robustness to the environment and the tissue’s model.

Representative work: Vision-based deformable trajectory control of soft tissue

Representative work: Simulation of soft tissue cutting

• Vision-based quadrotor control

As a passive sensor, cameras provide rich information at a low weight, small size and low power consumption, thus make them a suitable sensor for enabling quadrotor operation both in relatively narrow indoor environment and large, open spaces for exploring, searching and rescuing. Currently we are working on quadrotor visual serving; quadrotor SLAM; vision based target tracking and obstacle avoidance; vision based quadrotor control.

Applications：Inspection；Surveillance；Search and rescue； Disaster response；Fire fighting；