Energy efficiency in robotics software: a systematic literature review

This study presents a systematic literature review of software-level approaches to energy efficiency in robotics published from 2020 through 2024, updating and extending pre-2020 evidence. An automated-but-audited pipeline combined Google Scholar seeding, backward/forward snowballing, and large-language-model (LLM) assistance for screening and data extraction, with 10% human audits at each automated step and consensus-with-tie-breaks for full-text decisions. The final corpus comprises 79 peer-reviewed studies analyzed across application domain, metrics, evaluation type, energy models, major energy consumers, software technique families, and energy-quality trade-offs. Industrial settings dominate (31.6%), followed by exploration (25.3%). Motors/actuators are identified as the primary consumer in 68.4% of studies, with computing/controllers a distant second (13.9%). Simulation-only evaluations remain most common (51.9%), though hybrid evaluations are frequent (25.3%). Representational (physics-grounded) energy models predominate (87.3%). Motion and trajectory optimization is the leading technique family (69.6%), often paired with learning/prediction (40.5%) and computation allocation/scheduling (26.6%); power management/idle control (11.4%) and communication/data efficiency (3.8%) are comparatively underexplored. Reporting is heterogeneous: composite objectives that include energy are most common, while tasknormalized and performance-per-energy metrics appear less often, limiting cross-paper comparability. The review offers a minimal reporting checklist (e.g., total energy and average power plus a task-normalized metric and clear baselines) and highlights opportunities in cross-layer designs and in quantifying non-performance trade-offs (accuracy, stability). A replication package with code, prompts, and frozen datasets accompanies the review. Since 2020, building on gaps identified by Swanborn and Malavolta, the field has evolved. This study observed increased availability of energy-aware middleware, simulation platforms with energy-profiling capabilities, and design-time tools for estimating the energy impact of software decisions. Concurrently, emerging application domains such as swarm robotics, human-robot interaction, and adaptive mission planning have introduced new constraints and optimization challenges [3, 4] . This study responds to that need. This study presents a systematic literature review of studies published between 2020 and 2024 that address energy efficiency in robotics software. Building on the methodology developed by Swanborn and Malavolta, this study modernized and automated the search-and-screening pipeline (LLM-assisted) with manual audits, and applied both backward and forward snowballing to capture a comprehensive, current body of work. Ultimately, this study identified 79 peer-reviewed primary studies, which this study analyzed across application domain, used metrics, evaluation type, energy models, major energy consumers, software-level energy-saving techniques, and trade-offs with other system qualities. Methodological details (databases, search strings, screening procedures, automation and auditing) are provided in Section 2. A replication package with code, prompts, and frozen datasets (see Section 2) was also released. The contributions are threefold: C1 -Updated synthesis. An up-to-date synthesis of the field based on a substantially larger corpus (2020-2024), reflecting the evolving landscape of robotics and energy-aware software. C2 -Reproducible methodology. A reproducible, extensible methodology that leverages LLM-assisted screening and extraction with auditing mechanisms, accompanied by an open replication package. C3 -Categorized insights. A structured set of findings on how energy efficiency is addressed in robotic software today, highlighting dominant approaches, gaps, and promising directions for future work. The primary audience for this review includes both researchers seeking to extend energy-efficient robotics methodologies and practitioners looking for software strategies to reduce energy consumption in real deployments. This study is structured as follows: Section 2 details the methodology, encompassing search, screening, and data extraction. Section 3 presents the primary findings from the 79 selected studies. Section 4 explores trends, identifies gaps, discusses implications, and notes limitations. Finally, Section 5 concludes this study. Methodology This systematic literature review (SLR) extends Swanborn and Malavolta's 2020 study to literature published from January 1, 2020 (papers in the year 2020 that appeared in Swanborn and Malavolta's study were excluded manually at the start) through December 31, 2024. This study combines established SLR protocols with substantial automation to improve scalability, reproducibility, and transparency, following accepted guidelines for secondary studies in software engineering and robotics [5, 6] . A complete replicati