Akademik Veri Yönetim Sistemi
20th IEEE International Conference on Automation Science and Engineering, CASE 2024, Bari, İtalya, 28 Ağustos - 01 Eylül 2024, ss.1081-1087
Quadrupedal locomotion in environments without visual aid presents significant challenges in achieving automation capabilities. Often relying solely on IMU sensors and proprioceptive feedback from motors due to limited visual perception, understanding the robot's interaction with its environment becomes crucial for safe and efficient navigation. This paper presents a novel approach for estimating external forces applied to a quadrupedal robot's body, a critical aspect of its interaction with the environment. Our method integrates generalized momentum methods, contact constraints, and inertial measurement unit (IMU) data to accurately estimate these external forces. By using these diverse sources of information, the output of this work can enable navigation algorithms to enhance the robot's ability to navigate complex environments, improving stability and adaptability. Simulation studies and experimental validations demonstrate the effectiveness of the proposed method in real-world scenarios, highlighting its potential for advancing quadrupedal locomotion in unfamiliar terrains where visual aids are limited.