ADAPTIVE CONTROL MODELS FOR HUMAN-ROBOT COLLABORATIVE SYSTEMS IN LEAN WAREHOUSING: A CYBER-PHYSICAL APPROACH TO INDUSTRY 5
DOI:
https://doi.org/10.37547/Abstract
The transition from Industry 4.0 to Industry 5.0 necessitates a shift from purely autonomous automation to human-centric collaborative systems. While previous frameworks such as Dy- namic Value Stream Mapping (DVSM) successfully utilized IIoT data to quantify process waste, current collaborative environments suffer from "safety-induced latency"—where robots operate at suboptimal fixed speeds or stop entirely in the presence of humans. This paper proposes a novel Adaptive Control Model (ACM) that leverages a Cyber-Physical System (CPS) to synchronize robot trajectories with real-time human biometric and positional data. By modeling the warehouse floor as a dynamic Markov Decision Process, we implement a Deep Reinforcement Learning agent that adjusts robot velocity and proximity buffers based on predicted human intent and fatigue levels. Empirical validation through a high-frequency kitting simulation demonstrates that this adaptive approach reduces non-value-added (NVA) "hesitation time" by 22.0 % while maintaining ISO-compliant safety standards. This research provides a prescriptive roadmap for Lean Warehousing, transforming the value stream into a self-adjusting, ergonomic, and highly efficient collaborative ecosystem.
Downloads
References
[1]Breque, M., De Nul, L., and Petridis, A. (2021). Industry 5.0: Towards a sustainable, human- centric and resilient European industry. European Commission, Directorate-General for Re- search and Innovation.
[2]Romero, D., et al. (2016). The Operator 4.0: Human-Cyber-Physical Systems and Adaptive Automation. IFIP International Conference on Advances in Production Management Sys- tems.
[3]Müller, J. (2020). The Dawn of Industry 5.0: From Hardware-Centric to Human-Centric Manufacturing. International Journal of Precision Engineering and Manufacturing.
[4][User Name/Self-Reference] (2026). A Deep Reinforcement Learning (DRL) Framework for Dynamic Value Stream Mapping (DVSM) in Industry 4.0.
[5]Rother, M., and Shook, J. (2003). Learning to See: Value Stream Mapping to Add Value and Eliminate Muda. Lean Enterprise Institute.
[6]Liker, J. K. (2020). The Toyota Way: 14 Management Principles from the World’s Greatest Manufacturer. McGraw-Hill Education.
[7]Schulman, J., et al. (2017). Proximal Policy Optimization Algorithms. arXiv preprint arXiv:1707.06347.
[8]Haarnoja, T., et al. (2018). Soft Actor-Critic: Off-Policy Maximum Entropy Deep Rein- forcement Learning with a Stochastic Actor. International Conference on Machine Learning (ICML).
[9]Khatib, O. (1986). Real-Time Obstacle Avoidance for Manipulators and Mobile Robots. The International Journal of Robotics Research.
[10]ISO 10218-1:2011. Robots and robotic devices — Safety requirements for industrial robots.
[11]ISO/TS 15066:2016. Robots and robotic devices — Collaborative robots.
[12]Villani, V., et al. (2018). Survey on Human–Robot Collaboration in Industrial Settings: Safety, Intuitive Interfaces and Resilience. Mechatronics.
