Wireless-power mobile edge computing, SWIPT-MEC, priority-aware WPT scheduling and swarm resource allocation already solve important parts of the energy-management problem. The novelty of WPT-JCCO is not any one of those elements; it is a single swarm-supervisory feasible set that couples decisions which the three adjacent method classes normally separate. Each epoch-level action jointly selects the robot to charge and one of three physically distinct WPT modalities: far-field radio-frequency, resonant near-field and directional lightwave transfer, together with the SWIPT split, local/edge task placement, CPU frequency, bandwidth and transmit power. Relative to SWIPT-MEC, the formulation adds discrete recipient-modality selection with pose, alignment, blockage and dwell-dependent feasibility. Relative to conventional WPT scheduling, charging is not a separate priority or routing stage but is solved jointly with computation and radio allocation. Relative to swarm resource-allocation methods, energy replenishment is endogenous and an individual minimum-battery constraint protects the weakest robot. A fourth coupling makes the centrally generated resource vector admissible only when the complete sense-compute-actuate age fits the one-second supervisory epoch; otherwise a previously feasible or local-safe action is applied. Nonlinear harvesting, partial offloading, priority scoring and augmented-Lagrangian primal-dual updates are treated as established techniques. This paper derives the continuous block updates, keeps the WPT variables binary through candidate screening, and declares convergence only when stationarity, feasibility, merit-change and binary-hold tests are jointly satisfied. Normalised primal steps are safeguarded by backtracking, dual and penalty updates are bounded, and a local tracking bound plus divergence monitor delimit real-time operation without claiming global mixed-integer optimality or closed-loop motion stability. Numerical evaluation over a 20-robot swarm and 30 Monte Carlo runs shows that WPT-JCCO reduces net energy depletion by 23.8% relative to communication-computation optimisation with static WPT and by 49.7% relative to local-only execution, while increasing task success from 93.5% to 97.3%. A released common-trace comparison shows normalised-cost reductions of 11.1%, 11.3% and 5.8% relative to two-stage WPT+CCO, fixed-SWIPT dynamic offloading and an offline Q-learning scheduler. Convergence and one-factor-at-a-time sensitivity studies further examine swarm size, task load, WPT budget, bandwidth, edge capacity, mobility and channel margin. The headline values remain scoped to the nominal independent-task case; mode-specific RF, near-field and lightwave operating envelopes, robust pose/CSI, WPT-safety and task-DAG extensions are formulated but not presented as hardware-validated results.

This paper presents a joint communication–computation optimization (JCCO) framework for energy-efficient resource allocation in edge-assisted swarm robotics. The proposed framework jointly optimizes task offloading, bandwidth allocation, transmission power, and computational workload under latency and energy constraints. A distributed optimization strategy combined with a lightweight deep reinforcement learning controller enables adaptive and scalable decision making for resource-constrained robotic swarms. The simulation results demonstrate that the proposed method reduces the total swarm energy consumption by up to 41% while maintaining more than 99% deadline satisfaction across varying swarm sizes and communication conditions. The framework further achieves improved fairness, lower communication overhead, and efficient embedded deployment suitability for TinyML-enabled robotic platforms.

The Electric Vehicle Routing Problem (EVRP) is a key optimization challenge in autonomous and electric transportation. Unlike traditional routing, EVRP must consider battery constraints such as limited capacity and charging needs. Although routing methods have advanced, the integration of battery aging using realistic models remains underdeveloped. Addressing these dynamics is essential for improving long-term fleet efficiency. In this paper, we present a real-time, reconfigurable battery model that captures aging effects by updating key internal parameters based on the battery’s current State of Charge (SoC) and State of Health (SoH). Using this model, we formulate a multi-objective optimization problem and develop a genetic algorithm that balances energy efficiency, battery lifespan, and quality of service. Results show that incorporating aging-aware battery dynamics significantly extends battery life and reduces operational costs.

Outdoor surfaces equipped with underground liquid-based heating systems typically rely on simple linear rulebased controllers and sporadic manual overrides, leading to excessive energy use. To address this inefficiency, this paper presents an online Expert Advisor system that provides Quality of Service (QoS)-aware decision support for underground heating operations. Unlike prior offline models and autonomous modelpredictive controllers, the system delivers online decision support, integrating live environmental and operational data streams with a dynamic internal thermal model. The advisor employs an Extended Kalman Filter and Linear-Quadratic-Gaussian (EKF-LQG) control strategy to estimate temperature changes under varying weather conditions, enabling optimal heating recommendations that minimize energy costs while satisfying QoS requirements. By consolidating data acquisition, thermal state estimation, and actionable recommendations within a unified decision-support interface, the system renders the energy-quality trade-off explicit, enabling non-specialist operators to make datadriven decisions with greater confidence. Evaluated at Veritas Stadion in Turku, Finland, a retrospective analysis of one full heating season of operational data demonstrates a 35% reduction potential in heating energy consumption. Our work establishes a reproducible pathway toward human-in-the-loop optimization for large-scale hydronic underground heating infrastructures, deployable as an additional decision layer on top of existing control hardware.

Human Robot Collaboration in dynamic real world settings requires agents to perform high level intent inference without explicit rewards, rules, or corrective feedback. We present an Active Inference (AIF) framework formulated as a partially observable Markov decision process with discrete states. The method achieves self supervised goal recognition through minimization of variational free energy, enabling intrinsic inference of changing human intentions and autonomous action in non stationary environments.Intentions are treated as latent states. The agent integrates uncertain and noisy categorical observations that reflect human location, hand motions, and the fading of voice commands which serves as a proxy for changing trust. The model handles rapid learning of new transitions and effective reversal learning when context shifts abruptly. Simulation studies show that the approach maintains strong robustness. With both aleatoric and epistemic uncertainties the success rate is 79%. With only epistemic uncertainty the success rate reaches 86%.The method is further assessed on a physical Franka Emika Panda platform performing a dynamic handover task. Despite unmodeled noise and temporal ambiguity the robot reached approximately 68% accuracy. It shows that the AIF architecture can infer human intent from ambiguous and uncertain gestures in a manner similar to how people interpret each other, which in turn indicates that the framework can support adaptive and self supervised robotic partners.

This paper presents an adaptive simulation framework, implemented in NVIDIA Isaac Sim, for self-aware heterogeneous robotic systems. The proposed platform enables each robot to maintain an internal predictive model of its own embodiment and its interaction with the surrounding environment, while explicitly accounting for both mechanical actuation and onboard computational processes. By treating mechanical and computational energy as internal state variables, the framework enables runtime estimation of the energetic consequences of robot actions. This capability is demonstrated across diverse robot morphologies, including a mobile humanoid, a fixed-base manipulator, and an aerial robot operating in a shared warehouse environment. Experimental results show that internal computational processes constitute a substantial portion of the overall energy budget and are tightly coupled with mechanical action, highlighting the importance of incorporating internal resource state into self-regulatory decision-making. The framework provides a reusable platform for studying resource-aware self-regulation and energy-informed decision-making, with extensions toward multi-robot and swarm systems.

Energy-efficient coordination of robotic swarms requires effective integration of task scheduling, motion planning, and communication management, particularly in resource-constrained environments where computation and wireless communication compete for limited energy resources. Existing multi-robot approaches typically address these concerns in separate stages: task-allocation methods (e.g., market- and auction-based schemes) price assignments by distance, and computation-offloading methods decide execution placement after a route has been fixed. This paper’s specific contribution is to fold the execution-placement decision (local computation versus offloading to a peer) into the edge-relaxation step of an A* path search, using a composite cost whose communication term is derived from the instantaneous neighborhood of each node; routing and compute placement are therefore co-optimized within a single search rather than in decoupled stages. The framework is evaluated in simulation with a swarm of 25 robots against two decoupled baselines: a path-only planner that ignores workload and communication costs, and a workload-only scheduler that ignores travel and communication costs. Across 20 randomized trials, the proposed heuristic reduces total swarm energy consumption by approximately 22% relative to the path-only baseline and 9% relative to the workload-only baseline, shortens average task completion time by roughly 20%, and lowers the load imbalance factor from 6.7 (path-only) and 3.2 (workload-only) to 1.9. We report these gains for the tested configurations and delimit their scope: the search retains the asymptotic complexity of standard A*, but path optimality does not extend to the compute-placement decisions, which are locally greedy, and all results are obtained in simulation rather than on hardware.

This paper presents, to our knowledge, the first framework that jointly and provably optimizes communication and computation energy across an entire fleet of wirelessly connected mobile robots, rather than for a single device or under a fixed offloading policy as in prior work. Battery capacity limits the endurance of autonomous mobile robots, and on-board computation and radio communication increasingly rival locomotion in energy draw; across a fleet, the two are further coupled through shared uplink bandwidth and edge computing capacity. We formulate the joint selection of each robot’s task-offloading ratio, DVFS processor frequency, and transmit power, together with the fleet-wide allocation of bandwidth and edge capacity, subject to hard per-task deadlines. Closed-form inner solutions reduce each robot’s problem to a jointly convex program, coupled fleet-wide only through two linear resource constraints. We exploit this structure in JECCO-M, a distributed price-based algorithm that provably converges to the global fleet optimum while exchanging only a few scalars per iteration. A trajectory-conditioned channel-prediction extension handles robot mobility. Evaluated in simulations against optimization-based and learning-based baselines from the literature and on a physical three-robot testbed with embedded GPU compute, an IEEE 802.11ac uplink, and instrumented power rails, JECCO-M substantially reduces combined electronic energy while meeting all deadlines, and the measured hardware behavior tracks the analytical model closely. The results indicate that treating radio energy, processor energy, and shared edge resources as a single optimization domain is a practical route to extending the operating time of connected robot fleets.

The Electric Vehicle Routing Problem (EVRP) is a multi-objective, NP-hard optimization problem focused on efficiently managing various constraints involved in routing single or multiple electric vehicles. Recently, several evolutionary algorithms have been applied to EVRP, particularly those that incorporate battery State of Health (SoH) as a key constraint in the optimization process. However, most existing optimization algorithms treat SoH deterministically and do not account for the uncertainty inherent in its estimation. As a result, these methods may yield solutions that do not accurately reflect real-world battery conditions. To address this limitation, we propose a hybrid genetic algorithm that incorporates the stochastic process distribution of battery SoH estimation into the optimization procedure. Results from a lithium-ion battery-based fleet scenario show that incorporating uncertainty into the optimization improves battery SoH prediction accuracy by up to 70% compared to baseline methods.

In this paper, we propose a routing optimization framework for a mobile robot performing a specified mission, where the robot must navigate to a target location and return to the starting point while executing computation-intensive tasks. These tasks can be offloaded to a base station (BS) whenever communication coverage is available. Since robots are battery-powered, overall energy efficiency is a critical factor in routing decisions. Most existing routing methods focus on minimizing mechanical energy or onboard computational consumption, without considering the impact of computation offloading and communication energy on routing decisions. By explicitly considering computation offloading and its associated communication energy during path planning, routing decisions are influenced by the spatial placement of base stations, allowing the robot to adapt its path to exploit communication coverage. To this end, we formulate a routing problem that jointly optimizes mechanical, computational, and communication energy. Experimental results show that when communication energy is considered together with computation and motion energy through offloading-aware routing, the total energy consumption is reduced by approximately 13% on average compared to baseline methods that do not account for offloading.

Energy consumption is a critical concern for Internet of Things (IoT) platforms lacking abundant resources, particularly for swarm robotic systems that rely on numerous devices operating collaboratively over extended periods. This study presents a comprehensive design strategy for improving processing and communication to enhance system efficiency and reduce energy consumption. We incorporate energy harvesting (photovoltaic and RF), dynamic power management, and energy-efficient communication protocols (e.g., duty cycle, power control, data compression) into two complementary platforms built for swarm robotics: MCU-based nodes (TI MSP430 with LoRa transceiver), which serve as the experimental prototype for validating energy-aware communication, compression, and scheduling mechanisms; edge platforms (Jetson Nano and TX2), which are used for high-level power profiling and system-level evaluation, particularly for computation intensive workloads and comparative analysis. Our technique involves analyzing the device’s energy usage and harvesting processes, developing efficient communication protocols, and validating the system through simulations and hardware prototypes. Experimental results under outdoor and indoor conditions show that the device maintains an energy neutrality ratio well above unity, even with limited ambient energy. Key findings include significant reductions in energy per bit transmitted and reliable long-term operation. These insights pave the way for deploying swarms of autonomous IoT-based robots with minimal maintenance and maximal longevity.

The Internet of Things (IoT) renders swarm robotics possible, which makes tasks like surveillance, farming, and disaster response more efficient and flexible. The limited energy and processing power on board make things very difficult, especially since coordination requires a lot of communication. The goal of this paper is to lower the cost of communication in robotic swarms so that they last longer and make better use of resources. We offer a unified optimization framework that combines adaptive communication protocols with heuristic offloading algorithms. The main goal is to reduce communication energy, with computational cost being a secondary concern. The system changes the power levels and transmission rates based on how well the network is working. It uses a heuristic based on PSO to find the best balance between processing data locally and sending it to the cloud. Numerous evaluations on embedded platforms (NVIDIA Jetson Nano and TX2) demonstrate that the proposed method significantly conserves energy, reducing communication related energy consumption by approximately 25–35 percent relative to static schemes, while also mitigating CPU load fluctuations. Our results show that putting communication optimization first greatly improves swarm energy efficiency without hurting coordination performance.

Battery-powered mobile robots that rely on an edge server for perception spend energy in three places at once: the on-board processor, the radio front end and the drivetrain. These budgets are normally optimised separately, which is a mistake, as lowering the processor clock pushes work onto the wireless link, transmitting into a poor channel costs far more than waiting for a better one, and where the robot drives determines what the channel will be. We formulate the co-management of all three as a minimisation of long-run average energy for a fleet sharing an access point and subject to task deadlines, a power budget and a mission-progress constraint that forces every policy under comparison to cover the same ground. The resulting stochastic mixed-integer non-convex program is made tractable by a Lyapunov drift-plus-penalty argument that decomposes it into four per-slot subproblems: a square-root clock rule, a water-filling transmit-power rule with an explicit on/off test, a join-the-shorter-queue offloading split, and a short lookahead over admissible speeds. The policy needs no channel or workload statistics and attains an A per-task deadline mechanism, feasibility floors on the clock and transmit decisions, and closes the gap between queue-stability guarantees and individual task deadlines, which drift arguments alone do not bound. The policy needs no channel or workload statistics and attains an [O(1/V),O(V)] energy–delay tradeoff, stated under precisely qualified assumptions. In a per-task simulation study against six baselines, the policy reduced combined communication and computation power by 25% relative to the strongest deadline-compliant baseline (p<10−4) at equal mission progress, and was the only scheme to hold deadline violations below 0.5% across the full load range, where every baseline exceeded 16% at high load.

In this paper, we propose an integrated simulation framework for a mobile robot to analyze the dependencies among system components, particularly electrical power and battery models, rather than treating them independently. The framework captures the joint effects of mechanical power, computational power, and computational load offloading on total energy consumption and long-term battery health. It models mechanical, computational, and communication energy while accounting for their dependencies. As a case study, we apply the holistic framework to an energy-aware mission planning problem in which the robot travels to a target location and returns to its starting point. Along the planned path, computational tasks may be offloaded to base stations (BSs) when communication coverage is available. Because the robot operates on battery power, energy efficiency directly influences routing decisions and battery aging. The spatial distribution of BSs affects offloading opportunities, motion planning, resource allocation, and overall energy consumption, ultimately impacting battery degradation. Results from the case study demonstrate that the proposed integration of computational, mechanical, and communication energy models into routing optimization reduces total energy consumption by approximately 13% and battery degradation by 24% compared to approaches that optimize only mechanical and computational energy without considering load offloading.

The exponential growth of the Internet of Things (IoT) has led to an urgent need for highly energy-efficient communication strategies, especially for battery-powered or self-sustaining devices. In this work, we present a comprehensive framework for minimizing communication energy in IoT nodes operating in swarm robotic systems. We examine and integrate multiple low-power wireless technologies (BLE, LoRaWAN, MQTT, CoAP) with advanced Medium Access Control (MAC) protocols. We additionally propose adaptive scenarios leveraging both ambient energy harvesting and passive backscatter transmission. Our solution employs adaptive scheduling and dynamic transmission power management. Specifically, a Deep Q-Learning (DQL) agent dynamically adjusts transmission parameters based on the current energy condition. We develop analytical power consumption models incorporating equations for duty cycling, listening energy, and transmission energy. We then test the framework on a prototype IoT rover. Our experiments demonstrate that our optimizations extend device lifetime by up to 25% while maintaining rapid and reliable data transmission. The proposed methods substantially improve the balance between energy consumption and communication performance. This represents a significant advancement toward extending the operational lifetime of IoT and swarm robotic systems.

Lightweight industrial robots are increasingly deployed alongside humans to perform diverse and intelligent industrial tasks. A major concern with these robots is energy efficiency, driven by rising operational costs and environmental impacts. A growing contributor to energy use is the heavy computational workload of their electronic components. Although motion configurations and computational load are often interdependent, current state-of-the-art energy optimization methods tend to address them separately, focusing on individual consumption. In this work, we demonstrate that computational energy is comparable to mechanical energy and show how their dependency affects overall consumption in a Franka Emika Panda robot equipped with a multi-core processing system and two depth cameras. Building on this understanding, we propose a Bayesian approach for the joint optimization of mechanical motion and computational frequency in a robotic arm. Experiments show that the proposed method enables the Franka arm to reduce energy use by 3.7% in pick-and-place tasks and 6.2% in sorting tasks, compared to methods that optimize locomotion and computation separately.

Optimizing energy consumption is a critical challenge in autonomous mobile robotics, essential for extending battery life. Mechanical and computational components are the primary energy consumers, and studies show that dynamically co-managing their power usage—such as adjusting processing frequency relative to mechanical speed—significantly improves efficiency. This improvement is primarily due to the relationship between decision-making processes based on mechanical speed and computational workload. In this paper, we propose an agile reinforcement learning algorithm for dynamic co-management, tested on a rover equipped with a brushless motor, a Jetson TX2 processor, and an event-based camera. Our approach effectively addresses scalability and accuracy issues in prior methods, achieving energy efficiency improvements between 16.98% and 60.86% compared to the most efficient existing techniques.

Accurate estimation of battery State of Health (SOH) is crucial in the battery monitoring and management process. Several methods have been proposed to model and estimate battery aging dynamics, either formally, model-based or data-driven. One key challenge in SOH modeling is the generality of the SOH modeling approach, which requires consideration of inherent dependencies among the various multidisciplinary stress factors involved. In this paper, we present an end-to-end self-supervised approach based on Convolutional Auto-Encoders (CAEs) for learning informative intermediate features from battery measurable properties such as voltage, current and temperature. We then employ the learned features to estimate the change in battery SOH by a light-weight feed-forward neural network. The learned features represent essential information in battery dynamics and surpass the human-engineered features in terms of correlation with the target SOH characteristic. Utilizing these representative features, our SOH estimation model yields 58.7% and 45.0% average performance improvement on two large battery datasets compared to the state-of-the-art machine learning methods.

This paper proposes a structural simulation environment for cognitive systems that effectively models both bottom-up emergent needs derived from sensory data and top-down task specifications for actuator control. The proposed simulation structure is both scalable and explainable; needs emerge concurrently. The tasks triggered to fulfil these needs follow a hierarchical top-down approach. Experimental results demonstrate that the simulation environment accurately models the hunger and play needs of a fully autonomous pet robot, all within the overall goal of the robot’s survival.

Digital Twins (DTs) are increasingly applied for optimization of operations in logistics, healthcare, smart cities, and beyond. However, implementing high-fidelity DTs remains challenging in computationally intensive domains such as Computational Fluid Dynamics (CFD). While simplified models can facilitate real-time operation, they often lack physical fidelity. This article presents an open-source scalable software framework along with a case study of CFD-based digital twining on stormwater management. The presented framework enables online execution of CFD-based models by containerizing and integrating them into OpenShift platform, providing a two-way communication channel for simulation parameters and results. The framework is capable of dynamically scaling computing resources to run computationally-intensive CFD-models. In the case study, we present a novel CFD simulation model of a bioretention cell intended to reduce runoff volumes of urban stormwater. The simulation model, implemented in OpenFOAM, is then integrated into the presented software framework to create the DT. The framework source code, simulation model and the DT are made publicly available to promote future research.