Papers
RoboMaster Balancing Standard Robot
by Nicholas Chew, Leonardo Edgar, Xupravati William
This technical report documents the development and analysis of an advanced balancing infantry robot for the 2025 RoboMaster University Championship. The project addresses critical control challenges in dynamic battlefield environments through comprehensive system modeling and implementation of sophisticated control algorithms. The research team employed Newton-Euler mechanics for rigorous system dynamics analysis, developed kinematic models using Jacobian matrices for precise leg control, and implemented both classical Proportional-Integral-Derivative (PID) and optimal Linear Quadratic Regulator (LQR) control methodologies to achieve robust balancing and motion control. The carbon-fiber robot platform features innovative wheel-legged locomotion with jumping capabilities, enabling strategic navigation over elevated terrain obstacles. Extensive MATLAB/Simulink simulations validate the superior performance of the hybrid PID-LQR control architecture, demonstrating significant improvements in velocity tracking and angular stability compared to PID-only implementations. The study contributes valuable insights into bridging theoretical control methodologies with practical robotics applications, while establishing a foundation for future research incorporating Model Predictive Control (MPC) and reinforcement learning techniques to enhance autonomous operation in unpredictable competitive environments.
Development of Autonomous Tracking System in the RoboMaster University Championship (RMUC)
by Wong Ching Wee, Yee Zhan Xian Carlton
This technical report presents the development of an autonomous target acquisition and tracking system for deployment in the RoboMaster University Championship, specifically designed for the fully autonomous Sentry robot platform. The research addresses critical challenges in real-time computer vision and robotics control within dynamic competitive environments, where rapid target detection, precise pose estimation, and autonomous engagement capabilities are essential for strategic battlefield dominance. The system architecture integrates advanced hardware components including a high-performance Hikvision CS016-10UC camera with global shutter capability operating at 250fps, an Asus NUC 14 Pro+ mini PC for computational processing, and STM32-based microcontroller units for precise motor actuation control. The research methodology employs classical computer vision techniques including RGB color space segmentation, contour analysis with elliptical fitting, and Perspective-n-Point (PnP) algorithms for three-dimensional pose estimation, complemented by a lightweight Support Vector Machine classifier for symbol recognition and target prioritization. Experimental validation demonstrates superior power efficiency compared to deep learning approaches while maintaining robust performance metrics, including 92% classification accuracy and positional estimation errors within ±1cm at 2-meter distances. The modular system design utilizing Robot Operating System 2 (ROS2) ensures scalable integration with existing robotic platforms and provides a foundation for future enhancements incorporating Extended Kalman Filtering for predictive motion tracking and advanced optimization algorithms for improved computational efficiency in resource-constrained environments.
Development of Autonomous Movement and Navigation of a Sentry Robot for RMUC (Supercapacitor Buffer Implementation)
by Ng Weiliang
This engineering research presents the design and implementation of an advanced bidirectional power management system for autonomous robotic platforms competing in the RoboMaster University Championship, specifically addressing the critical challenge of operating high-performance robotic systems within stringent regulatory power constraints. The project develops a sophisticated supercapacitor-based energy buffer utilizing a four-switch bidirectional buck-boost converter topology controlled by the LT8708 integrated circuit, enabling dynamic power flow management between the primary 24V lithium-ion battery system and a 27V, 5F supercapacitor energy storage bank. The research encompasses comprehensive theoretical analysis of power conversion topologies, detailed component selection methodologies based on thermal and electrical stress calculations, and implementation of programmable bidirectional current limiting circuits utilizing external current sense amplifiers for precise load sharing control. Experimental validation demonstrates successful power regulation maintaining battery current below the 100W competition limit during transient load conditions exceeding 300W, achieving 81% round-trip energy efficiency while providing up to 2000 joules of supplemental energy storage. The system incorporates sophisticated analog control loops with real-time current monitoring and feedback mechanisms, enabling autonomous adaptation to varying load conditions during competitive scenarios. Future research directions include digital control system integration for enhanced stability optimization and development of intelligent power management algorithms to maximize energy utilization efficiency under dynamic operating conditions.