Evaluating Robot Program Performance with Power Consumption–Driven Metrics in Lightweight Industrial Robots

Juan Heredia, Emil Stubbe Kolvig-Raun, Sune Lundø Sørensen, Mikkel Baun Kjærgaard · 2025

Read Evaluating Robot Program Performance with Power Consumption–Driven Metrics in Lightweight Industrial Robots by Juan Heredia, Emil Stubbe Kolvig-Raun, Sune Lundø Sørensen, Mikkel Baun Kjærgaard (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-06295.txt.

Introduction

Juan Heredia and three co-authors grade robot programs by the power the robot draws in a 2025 paper. Code for factory robots is usually judged by processor stats, which say nothing about how the machine's body fares. A jerky program can waste energy and wear joints while looking fine on a chip chart.

Their frame reads the power curve of a run and turns it into plain scores for energy use, conversion, and steadiness, plus a wear score. Four programs for one machine-tending job were run on a standard arm. The scores split the programs cleanly and tied code choices to physical cost.

Catalog record

CREATORJuan Heredia, Emil Stubbe Kolvig-Raun, Sune Lundø Sørensen, Mikkel Baun Kjærgaard
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-06295

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Evaluating Robot Program Performance with Power Consumption–Driven Metrics in Lightweight Industrial Robots Juan Heredia, Emil Stubbe Kolvig-Raun, Sune Lundø Sørensen, Mikkel Baun Kjærgaard Abstract The code performance of industrial robots is typically analyzed through CPU metrics, which overlook the physical impact of code on robot behavior. This study introduces a novel framework for assessing robot program performance from an embodiment perspective by analyzing the robot’s electrical power profile. Our approach diverges from conventional CPU-based evaluations and instead leverages a suite of normalized metrics, namely, the energy utilization coefficient ( f U f_{U} ), the energy conversion metric ( f C f_{C} ), and the reliability coefficient ( f R f_{R} ), to capture how efficiently and reliably energy is used during task execution. Complementing these metrics, the established robot wear metric ( α \alpha ) provides further insight into long-term reliability. Our approach is demonstrated through an experimental case study in machine tending, comparing four programs with diverse strategies using a UR5e robot. The proposed metrics directly compare and categorize different robot programs, regardless of the specific task, by linking code performance to its physical manifestation through power consumption patterns. Our results reveal the strengths and weaknesses of each strategy, offering actionable insights for optimizing robot programming practices. Enhancing energy efficiency and reliability through this embodiment-centric approach not only improves individual robot performance but also supports broader industrial objectives such as sustainable manufacturing and cost reduction. I Introduction For over half a century, industrial robots have been essential in manufacturing. Recently, modern lightweight robots have transformed traditional industrial settings. These advanced robots, with built-in safety features, enable shared workspaces with humans. The International Federation of Robots (IFR) predicts a significant increase in their use [ 1 ] , highlighting their growing impact on future industrial operations. While the rise of lightweight industrial robots brings exciting opportunities, it also necessitates a comprehensive approach to performance evaluation. Traditional assessments have largely focused on either the mechanical capabilities, such as accuracy, repeatability, and speed [ 2 , 3 ] , or on computational efficiency, measured through CPU utilization, memory consumption, and latency [ 4 , 5 ] . However, these approaches overlook a critical dimension: the embodiment of the robot. In other words, while code performance is often judged by its computational footprint, it is equally important to understand how the execution of this code impacts the robot’s physical behavior, including its energy consumption and mechanical wear. Fig. 1: Assessing energy efficiency and reliability metrics in lightweight robots programs based on the robot power profile. This paper introduces a novel evaluation framework that assesses robot program performance from an embodiment perspective by analyzing the robot’s electrical power profile. Our approach quantifies the real-world impact in terms of energy efficiency and reliability of the robot program. Specifically, we introduce two normalized metrics for energy efficiency. The first metric, the energy utilization efficiency f_{U} , compares actual energy consumption to an idealized, loss-free model, measuring operational efficiency. The second metric, conversion efficiency f_{C} , assesses the conversion of electrical energy to mechanical work, indicating the proportion of energy effectively used. In addition to these energy metrics, we propose a Reliability Coefficient ( f_{R} ) based on the consistency of the robot’s power profile, which correlates power consumption patterns with performance stability. Complementing these, we employ an established robot wear metric ( \alpha ) [ 6 ] to quantify the mechanical stress imposed on the robot’s joints over time. We validate our integrated evaluation framework through a machine-tending case study using a UR5e robot, comparing four distinct programming strategies. This comprehensive analysis not only uncovers the strengths and weaknesses of each approach but also yields actionable insights for enhancing energy efficiency, reliability, and overall sustainability in industrial robotics. By linking code performance directly to physical outcomes, our work offers a deeper understanding of how programming choices affect energy dynamics and mechanical health, ultimately paving the way for more sustainable and reliable robotic operations in the future. The structure of the paper is as follows: Section II provides an overview of the state-of-the-art from the literature. Sections III and IV present the energy efficiency metrics and reliability metrics in detail. In Section V, we showcase a case study applying the framework to a machine-tending scenario. Section VI discusses the findings and reflections from the case study, and finally, Section VII concludes the article. II Related Work Assessing robot performance has long been a focal point in robotics research, particularly when it comes to energy efficiency and reliability metrics. Several studies have investigated absolute energy consumption as a key performance indicator in robotic manipulators [ 7 , 8 , 9 , 10 , 11 ] . These works employ data-driven models to analyze power usage under varying operational conditions and to optimize robot operations. While they have significantly advanced our understanding of energy consumption across different tasks, a notable limitation is the absence of normalized metrics that account for variations in task complexity and execution time. To address these gaps, researchers have proposed alternative metrics. For instance, metrics such as energy per distance traveled—or the “cost of transportation”—normalize energy usage by the work performed [ 12 ] . In addition, a benchmarking framework based on a set of metrics that compares the energy performance of various lightweight robots from different manufacturers was presented in [ 13 ] . Other works have introduced metrics focusing on computational efficiency in robotic software, such as “ J/m ,” “ J/hour ” [ 14 ] , performance-per-watt (“ FPS^{4}/W ”) for evaluating processing efficiency on different hardware platforms [ 15 ] , and the Energy Delay Product ( Et^{n} ) family [ 16 ] . While these approaches provide valuable insights, they focus either on mechanical or computational efficiency in isolation, limiting their applicability to holistic robot performance assessment. In parallel, research on robot reliability has traditionally relied on failure prediction methods such as fault-tree analysis, Markov models, and predictive maintenance techniques [ 17 ] . Recent studies, however, suggest that energy consumption profiles can serve as indicators of overall robot reliability and degradation [ 7 , 18 ] . This shift toward using power data for reliability estimation aligns with our approach, which introduces a reliability coefficient based on energy consistency. Additionally, [ 6 ] introduced a predictive model for estimating a robot’s end-of-life (EoL) using the stress metric ( \alpha ). While our method primarily focuses on runtime diagnostics and not EoL estimation, we recognize the value of this metric as an additional parameter. Incorporating it into our analysis allows for a more comprehensive assessment of the robot’s performance and reliability. Our research builds upon these foundations by proposing a comprehensive performance assessment framework that integrates energy efficiency, work performed, and reliability into a unified set of normalized metrics. Unlike previous studies that focus on single-dimensional measures, our approach enables fair comparisons between robotic programs regardless of task type. This directly supports the objectives outlined in the abstract: improving benchmarking, optimizing programming strategies, and enhancing energy-efficient robotics development. By addressing the limitations of existing methodologies, our proposed framework contributes to a more holistic evaluation of robot performance, paving the way for standardized sustainability metrics in industrial robotics. III Energy Efficiency

Read the complete text