A study published in Engineering explores a novel robot subset selection strategy for multi-user edge computing systems, designed to maximize the operational lifetime of robot swarms by exploiting correlation among distributed data sources in wireless networks. Conducted by researchers from the Institute for Communication Systems at the University of the Surrey, the work addresses a key challenge for 6G-enabled swarm robotics: excessive energy consumption from redundant data transmission, which shortens swarm operation when robots rely on battery power and edge computing for collaborative tasks.

The research defines a robot swarm as three or more cooperative robots with limited human control, a technology with applications in disaster recovery, agriculture and space exploration—all key use cases for 6G’s ultra-reliable low-latency communication, massive machine-type communication and integrated sensing and communication capabilities. The team notes that swarm lifetime is determined by the robot with the shortest battery life, as a single depleted robot halts overall swarm operation. Traditional multi-user edge computing processes all robot-sensed data individually, ignoring data correlation and leading to unnecessary communication overhead and rapid energy drain.

To solve this, the researchers model data correlation between spatially distributed robots as an undirected graph, introducing the concept of a robot subset: a group of robots whose combined sensed data is sufficient for the edge server to generate accurate computational outcomes, eliminating redundant transmissions. They formulate the lifetime maximization problem as a periodic subset selection challenge, then relax it into a graph partitioning problem and a subgraph-level vertex selection problem for computational feasibility. For additive white Gaussian noise (AWGN) channels, the team analyzes the theoretical upper bound of swarm lifetime and proposes two core algorithms: the least-degree iterative partitioning (LDIP) algorithm for graph partitioning and a final vertex search algorithm for subgraph-level vertex selection.

The study further adapts the framework for flat-fading channels, accounting for channel diversity and the base station’s limited channel estimation capabilities by integrating a max-min energy balance principle with the AWGN channel subset selection method. This adaptation reduces the number of robots requiring channel estimation while maintaining effective subset selection. Comprehensive simulations evaluate the approach across AWGN and independent and identically distributed Rayleigh fading channels, with baselines including conventional offloading (all robots transmit data) and a max-min algorithm for subset selection with limited channel estimation. Results show the proposed algorithms extend swarm lifetime by up to 650% compared to benchmark approaches, with performance gains driven by reduced communication energy consumption and balanced energy use across the swarm.

The research aligns with 6G edge intelligence goals, supporting reduced uplink congestion and fast coordination for real-world swarm robotics applications. The framework’s graph-based model and lightweight algorithms offer a scalable solution for energy-constrained robot swarms, providing a new perspective on resource optimization in multi-user edge computing for distributed robotic systems.

The paper “Robot Subset Selection-Based Multi-User Edge Computing for Swarm Lifetime Maximization with Correlated Data Sources,” is authored by Siqi Zhang, Yi Ma, Rahim Tafazolli. Full text of the open access paper: https://doi.org/10.1016/j.eng.2025.10.015. For more information about Engineering, visit the website at https://www.sciencedirect.com/journal/engineering.