Reconfigurable Intelligent Surface for Internet of Robotic Things

Wanli Ni, Ruyu Luo, Xinran Zhang, Peng Wang, Wen Wang, Hui Tian

This paper, titled "Reconfigurable Intelligent Surface for Internet of Robotic Things," explores how to turn the "uncontrollable" wireless environment into a programmable asset for multi-robot systems. By integrating Reconfigurable Intelligent Surfaces (RIS) into the Internet of Robotic Things (IoRT), the authors demonstrate how to simultaneously solve the headaches of signal blockage, limited battery life, and sensing inaccuracies.

TL;DR

The paper proposes an RIS-aided framework for the Internet of Robotic Things (IoRT) that optimizes communication, sensing, computation, and energy harvesting. By using advanced AI techniques like Federated Deep Reinforcement Learning, the authors show that RIS can significantly boost energy efficiency and sensing accuracy while helping robots navigate complex, obstacle-ridden environments.

What Problem Does This Paper Address?

The Internet of Robotic Things (IoRT) envisions fleets of robots working together, but they face a "triple threat" of constraints:

  1. Spectrum & Interference: As more robots connect, the airwaves get crowded.
  2. Physical Obstacles: Robots often operate in "cluttered" environments (warehouses, factories) where walls and machinery block high-frequency signals.
  3. Energy Scarcity: Mobile robots have limited battery life, and traditional wireless communication is power-hungry.

While traditional networks try to adapt to the environment, this paper argues we should change the environment using RIS—thin, smart surfaces that can reflect and steer electromagnetic waves toward moving robots, effectively "bending" signals around corners.

Methods

The authors break down the problem into three distinct case studies, applying different mathematical and AI-driven tools to each:

  • Federated Deep Reinforcement Learning (FDRL): Used for joint communication and trajectory design. This two-stage approach uses a "Global" stage to configure the RIS and a "Local" stage for individual robot path planning. The federated nature allows robots to learn from each other without sharing raw data, improving scalability and privacy.
  • Multi-Objective Optimization (MOO): Applied to Integrated Sensing and Communication (ISAC). Since sensing a target and transmitting data often compete for the same resources, the authors use a weighted Tchebycheff-based method to find the "Pareto front"—the sweet spot where both sensing accuracy and data rates are maximized.
  • Double-Agent Deep Reinforcement Learning (DRL): Used for "Over-the-Air Computation" (AirComp) and wireless charging. This handles the uncertainty of "channel aging" (where the signal changes faster than the robot can track it) by having two agents cooperate to optimize energy beamforming and data aggregation.

Datasets

The paper relies on numerical simulations rather than static public datasets. The authors modeled a multi-robot environment with the following parameters:

  • Network Setup: A base station (BS) serving multiple mobile robots (K) in a 30m x 30m area cluttered with obstacles.
  • Hardware Specs: RIS units with varying elements (20 to 40), and multi-antenna base stations (up to 32 antennas).
  • Channel Models: They accounted for both Far-field (planar waves) and Near-field (spherical waves) propagation, incorporating path loss exponents of 3.5 and noise power of -100 dBm.

Key Insights and Ideas

  • The Environment is Now a Variable: Traditionally, the wireless channel is a "given" that engineers must endure. RIS turns the channel into a software-defined parameter that can be optimized alongside the robot's physical path.
  • Trajectory-Communication Synergy: A robot’s path shouldn't just be the shortest distance between two points; it should be the path that maintains the best "energy beam" from an RIS to keep the robot charged and connected.
  • Scalability via Federation: Centralized AI struggles as the number of robots grows. By using Federated Learning, the system reduces the computational load on the central base station while allowing robots to adapt to local obstacles.
  • ISAC Trade-offs: The paper quantifies the fundamental trade-off between sensing and communication. It shows that coordinated beamforming across multiple base stations, aided by RIS, can drastically reduce sensing errors that a single base station couldn't overcome.

Key Results

  • Energy Efficiency: The proposed FDRL algorithm significantly outperformed standard Orthogonal Multiple Access (OMA) and Non-Orthogonal Multiple Access (NOMA) setups, showing that RIS-aided path planning is more energy-efficient than simply boosting transmit power.
  • Sensing Gains: In multi-cell networks, the coordinated RIS approach achieved a gain ratio of up to 20x in sensing/communication performance as the number of antennas increased.
  • Computation Accuracy: For AirComp (aggregating data in the air), the Double-Agent DRL reduced the Mean Square Error (MSE) by 57.6% compared to single-agent approaches, even when the robots had imperfect knowledge of the signal environment.

Limitations and Future Work

  • CSI Acquisition: The authors acknowledge that getting perfect Channel State Information (CSI) for a moving robot and a reflecting surface is extremely difficult in practice.
  • Hardware Realism: The simulations assume relatively ideal RIS behavior; real-world hardware has discrete phase shifts and "quantization errors" that might degrade performance.
  • Future Directions: The authors suggest moving toward "Embodied Intelligence," where Large Language Models (LLMs) and multimodal AI help robots collaborate more naturally with humans while the RIS handles the underlying connectivity fabric. They also highlight the need for Ultra-Reliable Low-Latency Communication (URLLC) for safety-critical robotic tasks.

Summary updated

Abstract

With the rapid development of artificial intelligence, robotics, and Internet of Things, multi-robot systems are progressively acquiring human-like environmental perception and understanding capabilities, empowering them to complete complex tasks through autonomous decision-making and interaction. However, the Internet of Robotic Things (IoRT) faces significant challenges in terms of spectrum resources, sensing accuracy, communication latency, and energy supply. To address these issues, a reconfigurable intelligent surface (RIS)-aided IoRT network is proposed to enhance the overall performance of robotic communication, sensing, computation, and energy harvesting. In the case studies, by jointly optimizing parameters such as transceiver beamforming, robot trajectories, and RIS coefficients, solutions based on multi-agent deep reinforcement learning and multi-objective optimization are proposed to solve problems such as beamforming design, path planning, target sensing, and data aggregation. Numerical results are provided to demonstrate the effectiveness of proposed solutions in improve communication quality, sensing accuracy, computation error, and energy efficiency of RIS-aided IoRT networks.
BibTeX
@article{ni2025reconfigurable,
  author = {Wanli Ni and Ruyu Luo and Xinran Zhang and Peng Wang and Wen Wang and Hui Tian},
  title = {{Reconfigurable Intelligent Surface for Internet of Robotic Things}},
  journal = {IEEE Internet of Things Magazine},
  year = {2025},
  doi = {10.1109/iotm.001.2400208},
}
Figure