Design of a Low-Cost Indoor Navigation System for Food Delivery Robot Based on Multi-Sensor Information Fusion

Area of Science:

• Robotics and Automation
• Sensor Fusion
• Navigation Systems

Background:

• The restaurant industry faces labor shortages, increasing interest in automation.
• Current meal delivery robots often rely on magnetic navigation, requiring costly infrastructure changes.
• Existing systems lack flexibility and are difficult to adapt to dynamic restaurant layouts.

Purpose of the Study:

• To develop a cost-effective and flexible navigation system for non-rail meal delivery robots.
• To overcome the limitations of magnetic navigation systems in restaurant environments.
• To improve the positioning accuracy and maneuverability of autonomous delivery robots.

Main Methods:

• Multisource information fusion combining ultra-wide band (UWB) positioning with odometer and low-cost inertial sensors (gyroscope and accelerometer).
• Integration of an electronic compass for enhanced steering control and path following.
• Application of the extended Kalman filter (EKF) to fuse UWB data with track estimation algorithms.

Main Results:

• Achieved precise positioning for a non-rail meal delivery robot with an average error of approximately 15 cm.
• Demonstrated flexible and cost-effective navigation along fixed orbits with steering control.
• The proposed multisource fusion approach proved effective in overcoming magnetic navigation limitations.

Conclusions:

• The developed multisource information fusion system offers a viable and adaptable solution for autonomous meal delivery robots.
• This technology presents significant potential for commercial application in the food service industry, addressing labor shortages and operational costs.
• The system provides a flexible, cost-effective alternative to traditional magnetic navigation, enhancing robot utility in dynamic environments.