Offload or Overload: A Platform Measurement Study of Mobile Robotic Manipulation Workloads

2026-03-18

AI Summary

This study addresses the challenges of deploying compute-intensive foundation models for mobile robotic manipulation across onboard, edge, and cloud platforms, where trade-offs arise among deployment feasibility, energy consumption, and network latency. For the first time, it systematically quantifies the execution characteristics of end-to-end manipulation workloads on heterogeneous GPU platforms—spanning onboard, edge, and cloud—using a real robotic system and multidimensional metrics including latency, energy, accuracy, and bandwidth. The findings reveal that small onboard GPUs cannot sustain full-task execution, while larger onboard GPUs significantly reduce operational endurance. Pure cloud offloading is constrained by limited bandwidth and degrades task accuracy. Furthermore, the work identifies viable scenarios for collaborative inference among robot swarms and exposes critical bottlenecks in shared resource utilization, offering actionable insights for the design of distributed robotic systems.

Technology Category

• Intelligent Robots: Manipulation
• Machine Learning: Learning on the Edge & Model Compression
• Humans and AI: Other Foundations of Human Computation & AI

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Mobile robotic manipulation--the ability of robots to navigate spaces and interact with objects--is a core capability of physical AI. Foundation models have led to breakthroughs in their performance, but at a significant computational cost. We present the first measurement study of mobile robotic manipulation workloads across onboard, edge, and cloud GPU platforms. We find that the full workload stack is infeasible to run on smaller onboard GPUs, while larger onboard GPUs drain robot batteries several hours faster. Offloading alleviates these constraints but introduces its own challenges, as additional network latency degrades task accuracy, and the bandwidth requirement makes naive cloud offloading impractical. Finally, we quantify opportunities and pitfalls of sharing compute across robot fleets. We believe our measurement study will be crucial to designing inference systems for mobile robots.

Problem

• mobile robotic manipulation
• computational offloading
• onboard GPU
• network latency
• compute sharing

Innovation

• mobile robotic manipulation
• workload measurement
• compute offloading
• foundation models
• robot fleet sharing