| A Review of Advancements in Collision Avoidance for Autonomous Robots |
| کد مقاله : 1218-NAEC (R1) |
| نویسندگان |
|
Alexandria Wampamba *1، Mansour Hakim-Elahi2، Massoud Masih Tehrani2 1Student 2Professor |
| چکیده مقاله |
| The quest for affordable autonomy has driven significant innovation in collision avoidance for autonomous robots between 2020 and 2025. The primary objective is to achieve robust navigation in dynamic environments using affordable hardware, thereby democratizing access to reliable autonomy. The subject matter encompasses bio-inspired algorithms, efficient sensor fusion, and lightweight spatial mapping techniques. Methods of investigation include the implementation of monocular vision with deep learning models like TinyYOLO, radar-ultrasonic fusion for all-weather reliability, and reinforcement learning for adaptive path planning. Noteworthy achievements include the successful deployment of hippocampal-inspired memory models enabling multi-resolution mapping in resource-constrained systems, alongside parsimonious perception strategies that reduce computational load by over 60%. These innovations have enabled robots as small as 70mm to navigate complex settings using only microcontroller-level hardware. The integration of neuromorphic computing and swarm intelligence emerges as a promising direction for future ultra-low-power navigation. Collectively, these developments bridge the performance gap between high-cost and affordable systems, making advanced collision avoidance feasible for widespread applications in agriculture, logistics, and consumer robotics |
| کلیدواژه ها |
| Low-cost collision avoidance, Spatial memory, sensor fusion, autonomous robots |
| وضعیت: پذیرفته شده |
