Cooperative Robots for Object Sorting and Distribution

Description

Research Problem Modern warehouses increasingly rely on autonomous mobile robots to improve efficiency and reduce operational costs. While many multi-robot systems successfully perform object transportation, efficient cooperation remains a significant research challenge. Existing approaches often use simple nearest-task or first-available assignment strategies that fail to consider future interactions between robots, workspace congestion, or the effect of current decisions on subsequent tasks. As a result, robots may experience unnecessary travel, waiting time, unbalanced workloads, and reduced overall throughput. This project investigates how intelligent cooperative decision-making can improve the efficiency of multi-robot warehouse sorting using low-cost robotic platforms. The central research question is whether incorporating predictive task allocation and coordinated path planning can significantly reduce total task completion time while maintaining low computational complexity. Research Aim The aim of this research is to design, implement, and experimentally evaluate an intelligent cooperative multi-robot warehouse sorting system that dynamically allocates tasks and coordinates robot actions to improve operational efficiency. Research Objectives The research objectives are to: Develop a vision-based framework for detecting coloured boxes and localising multiple mobile robots. Design a cooperative task allocation algorithm that assigns tasks based on estimated execution cost rather than simple distance. Investigate scheduling strategies that consider future accessibility, robot availability, and potential waiting time. Develop coordinated navigation methods that minimise movement conflicts and idle periods. Experimentally evaluate the proposed algorithms using a low-cost robotic warehouse prototype. Proposed Research Methodology The research will develop a prototype consisting of two autonomous mobile robots operating within a structured warehouse environment. A single overhead camera connected to a Raspberry Pi 5 will provide global visual information, including robot localisation, coloured box detection, and workspace mapping. Each robot will execute assigned pickup and delivery tasks while continuously reporting its status through wireless communication. The primary contribution is an adaptive scheduling algorithm that estimates the overall cost of each possible assignment using multiple factors, including travel distance, turning time, pickup duration, delivery time, expected waiting time, and the influence of current decisions on future task accessibility. The scheduler continuously updates assignments according to real-time observations, allowing the system to adapt to changing workspace conditions. Research Contribution Unlike conventional task allocation methods that optimise only the immediate task, the proposed approach introduces limited lookahead planning to evaluate the future consequences of current assignments. The research investigates whether considering future accessibility and robot interactions can improve overall warehouse performance without introducing excessive computational overhead. The study contributes an experimentally validated cooperative scheduling framework suitable for low-cost robotic platforms that combines computer vision, autonomous navigation, and intelligent decision-making. Experimental Evaluation The proposed approach will be compared with conventional nearest-task and greedy scheduling algorithms. Performance will be evaluated using several quantitative metrics, including total completion time, average robot idle time, travel distance, workload balance, communication latency, planning overhead, and sorting accuracy. Statistical analysis will determine whether the proposed cooperative strategy provides significant performance improvements while maintaining practical real-time operation.