[Paper Review] Review of Autonomous Mobile Robots for the Warehouse Environment

Russell Keith, Hung Manh La|arXiv (Cornell University)|Jun 12, 2024

Advanced Manufacturing and Logistics Optimization4 citations

TL;DR

This paper reviews autonomous mobile robots (AMRs) in warehouse environments, focusing on hardware, robotic control (localization, path planning, AI), and system-level control (scheduling, resource management, human-robot collaboration). It highlights how AMRs reduce motion waste, improve efficiency, and enable scalable, decentralized operations through sensor fusion, advanced algorithms, and optimized warehouse layouts.

ABSTRACT

Autonomous mobile robots (AMRs) have been a rapidly expanding research topic for the past decade. Unlike their counterpart, the automated guided vehicle (AGV), AMRs can make decisions and do not need any previously installed infrastructure to navigate. Recent technological developments in hardware and software have made them more feasible, especially in warehouse environments. Traditionally, most wasted warehouse expenses come from the logistics of moving material from one point to another, and is exhaustive for humans to continuously walk those distances while carrying a load. Here, AMRs can help by working with humans to cut down the time and effort of these repetitive tasks, improving performance and reducing the fatigue of their human collaborators. This literature review covers the recent developments in AMR technology including hardware, robotic control, and system control. This paper also discusses examples of current AMR producers, their robots, and the software that is used to control them. We conclude with future research topics and where we see AMRs developing in the warehouse environment.

Motivation & Objective

• To analyze recent advancements in AMR technology for warehouse automation.
• To identify key challenges in AMR deployment, including path planning, battery management, and human-robot collaboration.
• To evaluate the role of AI, sensor fusion, and decentralized control in improving AMR scalability and performance.
• To examine warehouse layout and zoning strategies that optimize AMR fleet efficiency.
• To identify gaps in current research, particularly regarding human factors and dynamic work allocation in AMR systems.

Proposed method

• Conducted a literature review using IEEE Xplore and Google Scholar, filtering for papers from 2013 to 2024 with keywords like 'AMR', 'warehouse', 'path planning', 'scheduling', and 'decentralized'.
• Categorized studies into hardware (sensors, processors, batteries), robotic control (localization, AI, path planning), and system control (resource management, scheduling, human-robot picking methods).
• Evaluated real-world AMR implementations from companies like Amazon (Kiva robots) and analyzed their software and control architectures.
• Reviewed optimization techniques such as simulated annealing, genetic algorithms, and hybrid methods (e.g., NSGA-II with tabu search) for layout and routing design.
• Assessed performance metrics including lead time, distance traveled, tardiness, and operational efficiency in AMR scheduling and routing.
• Explored non-traditional warehouse layouts using algorithm-generated designs that deviate from standard grid patterns to improve throughput.

Figure 1: Human-robot collaboration in the warehouse [ 5 ]

Experimental results

Research questions

• RQ1 How do modern AMRs achieve navigation and localization without pre-installed infrastructure, and what sensor technologies enable this?
• RQ2 What role do AI techniques like deep reinforcement learning and genetic algorithms play in optimizing AMR path planning and scheduling?
• RQ3 How can decentralized scheduling and dynamic zoning improve scalability and reduce congestion in large-scale AMR fleets?
• RQ4 What are the key performance trade-offs between traditional warehouse layouts and non-conventional, algorithm-optimized layouts?
• RQ5 Why are human factors and ergonomic considerations underrepresented in current AMR-based order picking systems?

Key findings

• AMRs significantly reduce motion waste in warehouses by minimizing human walking distances, improving efficiency and reducing fatigue in human collaborators.
• Sensor fusion using LiDAR, 2D/3D cameras, and inertial sensors enables accurate real-time localization and environment mapping without fixed infrastructure.
• Decentralized scheduling and dynamic zoning strategies show promise for scaling AMR fleets beyond 10 robots, though few studies address hundreds of robots in real-world settings.
• Non-traditional warehouse layouts generated by optimization algorithms (e.g., using simulated annealing or hybrid metaheuristics) can increase throughput by eliminating fixed aisles and optimizing item placement.
• Genetic algorithms and hybrid methods like NSGA-II with tabu search effectively reduce material handling costs by optimizing AGV/AMR path flows in multi-row warehouse configurations.
• Despite progress, human factors such as workload distribution, fatigue, and ergonomics remain underexplored in AMR-integrated order picking systems.

Figure 2: Amazon’s Kiva robot [ 6 ]