Adaptive Trajectory Tracking of Nonholonomic Mobile Robots Using Vision-Based Position and Velocity Estimation

Area of Science:

• Robotics
• Control Systems
• Computer Vision

Background:

• Trajectory tracking control (TC) for nonholonomic mobile robots (NMRs) is challenging without global positioning systems.
• Accurate robot localization using only onboard sensors remains a significant hurdle in autonomous navigation.

Purpose of the Study:

• To propose a novel adaptive trajectory tracking control method for NMRs that does not require position, orientation, or velocity measurements.
• To develop an algorithm for online estimation of robot position and velocity using visual feedback from an omnidirectional camera.

Main Methods:

• An adaptive control strategy is designed for trajectory tracking.
• A novel algorithm estimates robot pose and velocity in real-time using an omnidirectional camera.
• Theoretical analysis proves the asymptotic convergence of tracking errors.

Main Results:

• The proposed method enables trajectory tracking control for NMRs without external localization.
• Online estimation of robot state variables (position, velocity) is achieved through visual feedback.
• Experimental validation on a wheeled NMR confirms the system's feasibility and effectiveness.

Conclusions:

• The developed adaptive TC method successfully addresses the open problem of NMR control without GPS.
• The novel visual feedback-based estimation algorithm provides accurate online state information.
• This research advances autonomous navigation capabilities for mobile robots in GPS-denied environments.