We introduce ET-Former, a novel end-to-end algorithm for semantic scene completion using a single monocular camera. Our approach generates a semantic occupancy map from single RGB observation while simultaneously providing uncertainty estimates for semantic predictions. By designing a triplane-based deformable attention mechanism, our approach improves geometric understanding of the scene than other SOTA approaches and reduces noise in semantic predictions. Additionally, through the use of a Conditional Variational AutoEncoder (CVAE), we estimate the uncertainties of these predictions. The generated semantic and uncertainty maps will help formulate navigation strategies that facilitate safe and permissible decision making in the future. Evaluated on the Semantic-KITTI dataset, ET-Former achieves the highest Intersection over Union (IoU) and mean IoU (mIoU) scores while maintaining the lowest GPU memory usage, surpassing state-of-the-art (SOTA) methods. It improves the SOTA scores of IoU from 44.71 to 51.49 and mIoU from 15.04 to 16.30 on SeamnticKITTI test, with a notably low training memory consumption of 10.9 GB, achieving at least a 25% reduction compared to previous methods. Project page: https://github.com/amazon-science/ET-Former.

The measurement accuracy of electric energy meters (EEMs) is crucial for respecting the fairness of the electricity market and the justice of electric energy settlement. However, the basic error (BE) of the measurement suffers from external stress, especially under extreme environments. It is challenging to evaluate the degradation trend of EEMs under multi-stress. To address this issue, this paper proposes modeling and evaluating methods for precise degradation analysis of EEMs under multi-stress. First, an optimized k-nearest neighbor (OKNN) is proposed to correct the outliers, where the stress-related weight and weighting factor are designed. Then, a varied-bias wiener process with hierarchical Bayesian (VWHB) model is introduced to evaluate and explore the impact of stress on the BE. The multi-stress, as well as the BE, are fused to parameterize this impact. Integrating the OKNN and VWHB, a degradation analysis framework is further proposed to model the data and conduct the degradation analysis. Extensive experiments on field data collected from the typical operating environment lab demonstrated that the proposed degradation analysis framework reveals superior performance than some state-of-art approaches. The root of the mean of the square of errors and the mean of absolute value of errors of OKNN-VWHB are the lowest with 0.0356 and 0.0281, respectively, for all the EEMs from three different companies.

When a mobile robot autonomously explores an indoor space to produce a localization and navigation map, it is important to create both a stable pose graph and a high-quality occupancy map that covers all the navigable areas. In this work, we propose a novel probabilistic active loop closure framework which attempts to maximally reduce pose graph uncertainty during exploration and improves occupancy map quality. We calculate a probabilistic reward of getting a loop closure at any pose on a pose graph, which considers both how much pose graph uncertainty would be reduced by getting a loop closure there, and the robot’s travel cost to navigate to that pose. By choosing poses that provide the largest rewards, we can maximally reduce pose graph uncertainty while avoiding long travel times. The effectiveness of the method is illustrated through on-device testing in various floor plans.

This article presents a novel approach to physical-displacement-based power grid measuring via an intensity-modulated fiber-optic sensor (IMFOS). An IMFOS utilizes one fiber to transmit the intensity modulated light from its electro-optic controller to a fiber-optic probe. The power grid voltage and current can induce physical displacements in transducers via the piezoelectric effect and the Lorentz law, respectively, which then result in a distance change between the optical probe and the reflective surface of the transducers. In parallel, multiple fibers are used to collect the reflective light for electro-optic conversion. A National-Instruments-based characterization platform is set up for performance evaluation. The testing result demonstrates that the IMFOS is immune to the inherent dc and low-frequency saturation issues prevalent in conventional potential and current transformers. Finally, the IMFOS is implemented in a universal grid analyzer to illustrate its applicability for phasor estimation in actual power grids.

Neural network controllers have become popular in control tasks thanks to their flexibility and expressivity. Stability is a crucial property for safety-critical dynamical systems, while stabilization of partially observed systems, in many cases, requires controllers to retain and process long-term memories of the past. We consider the important class of recurrent neural networks (RNN) as dynamic controllers for nonlinear uncertain partially-observed systems, and derive convex stability conditions based on integral quadratic constraints, S-lemma and sequential convexification. To ensure stability during the learning and control process, we propose a projected policy gradient method that iteratively enforces the stability conditions in the reparametrized space taking advantage of mild additional information on system dynamics. Numerical experiments show that our method learns stabilizing controllers with fewer samples and achieves higher final performance compared with policy gradient.

Charging station that incorporates renewable energy resource and energy storage is a promising solution to meet the growing charging demand of electric vehicles (EVs) without the need to expand the distribution network. The aggregation of multiple energy resources and EVs requires an efficient and flexible energy management strategy. This article presents a two-stage scheme to solve the power allocation and charging coordination of plugged-in EVs. Game-theory-based control is utilized to address the interaction among different components for respecting their individual preferences. The first stage determines the power allocation of photovoltaic, battery, and the grid as well as total charging power for EVs. In the second stage, charging power dispatch among individual EVs is coordinated based on the available total charging power determined in the first stage. As a result, the two energy management problems of charging station are addressed sequentially. The proposed solution is validated via simulation and experiment, and the comparisons with benchmarks show its advantages.

This article proposes a charging management of electric vehicles (EVs) with a newly presented EV social contribution. The EV charging problem is represented by a generalized Nash equilibrium game where each individual EV tries to minimize its charging cost while satisfying its own charging requirements and respecting the charging facility constraints. The individual EV features a social behavior to potentially contribute in shifting its charging schedule from specific intervals that have insufficient charging power. This shift in the EV schedule will allow more charging power to other EVs that admit stricter charging requirements, i.e., intervals and demands. In this way, the contributed EVs socially help others in reducing their charging costs, which is particularly important during the overload cases in the system. The proposed solution is reached iteratively in a distributed way utilizing the consensus network and found based on the receding horizon optimization framework. Both simulation and experimental results demonstrate the effectiveness and correctness of the proposed social contribution in the charging management for reducing the charging cost of EVs.

A hierarchical distributed energy management for multiple photovoltaic (PV) based electric vehicle (EV) charging stations (PV-CSs) is proposed and analyzed in this paper. In the station level, PV-CSs are modelled as independent players with objectives to stabilize their average available capacity (AAC) of the storage battery tank. Meanwhile, in EV level, EV owners are modeled as players with objectives to maximize their charging power. Then a two level power distribution game is utilized to model the power distribution problem in both station and EV level. Through utilizing a consensus network based learning algorithm, a cooperative and a generalized Stackelberg equilibrium are achieved in station and EV level through a distributed fashion. One case studies, i.e., two station case, is implemented in simulation to verify the performance and effectiveness of the proposed strategy. The simulation results show that the proposed energy management has an excellent performance in both cases and comparing against stations without management.

This paper discusses decentralized real-time energy management for multiple-source hybrid energy systems (HESs), which adapts to the sudden change in system configuration, such as due to failure of certain devices. An engine-generator/battery/ultracapacitor (UC) HES is chosen as a case study facilitating the following theoretical discussion. The energy management problem is first modeled as a noncooperative game, in which the different preferences of the energy sources (engine-generator, battery pack, and UC pack) in actual operation are quantified through their individual utility functions. The Nash equilibrium is iteratively reached at each control instant via a learning algorithm. Under the game theory based control, each source or player tends to maximize its own preference. However, its satisfaction level also depends on decisions of others. This real-time interaction in decision making provides the proposed energy management a capability to autonomously adapt to the reconfigured HES. A tuning procedure of weight coefficients in the utility functions also helps to further improve the adaptiveness of the decentralized energy management. Both the simulation and real-time implementation show that the game theory based energy management strategy has a comparable performance to the classical centralized benchmarking strategy. Meanwhile, the decentralized strategy demonstrates an obvious flexibility handling the cases when the configuration of the HES varies both statically and dynamically.

This paper introduces a distributed energy management for islanded DC microgrids with electric vehicles (EVs) penetration. The energy management problem is divided into two stages. The first, named as filtering stage, is controlled through wavelet filter using the ultracapacitor. The second one is modeled as a noncooperative stackelberg game, where the battery energy storage system is designed as a leader and EVs as followers. The preference of the battery energy storage system is designed to extend its cycle life while for each EV's preference is to maximize its charging power. The consensus network is utilised to reach the Nash equilibrium within the followers at each control instant. The simulation results show the power distribution among the system's units throughout one day. Moreover, comparable outcomes with the centralized method have been illustrated under two available power scenarios i.e., two weather profiles.

A distributed energy management for a photovoltaic (PV) based electric vehicles (EVs) charging station is proposed and discussed in this paper. The PV-based charging station and EV owners are modeled as independent players with different preferences where the preferences of the players are described through linear and logarithmic functions. Then, a noncooperative power distribution game is set up and a generalized stackelberg equilibrium is found at any control instant. Through utilizing the consensus network based learning algorithm, the charging power of EVs can be updated in a distributed way. The single stage, dynamic case study and scalability analysis in the simulation show that the proposed energy management can be well applied in the PV-based charging station and have an excellent performance.

This paper discusses a decentralized energy management for an engine-generator/battery/ultracapacitor (UC) hybrid energy system with fault tolerance analysis. The energy management problem among the energy suppliers and the load is formed into a non-cooperative power distribution game where the engine-generator, the battery pack, the UC pack, and the load are modeled as independent and related players. Each player has an unique objective, i.e., reducing fuel consumption, prolonging battery cycle life, maintaining UC state of charge and satisfying the load demands, represented by different second order polynomial function based utility functions. In this game, a Nash equilibrium is reached at each control instant to give a balanced solution among players. The weight coefficients in the utility functions can be determined through the Pareto optimal solution a multi-objective genetic algorithm. The fault tolerance analysis based simulation shows that the proposed energy management has a flexible and reconfigurable performance under six different case studies.

This paper provides a two-level energy management strategy for a fuel cell-battery-ultracapacitor (UC) hybrid system. In the proposed strategy, the battery and UC packs are seen as an energy storage system (ESS) at the first level and the equivalent consumption minimization strategy is used to distribute load power between this ESS and the fuel cell system. The penalty factor is tuned based on estimated average load power and SOC of the ESS. At the second level, the power distribution between the battery and UC packs is determined using the equivalent series resistance-based control strategy to minimize the energy loss. Then, the performance of the proposed two-level energy management strategy is analyzed in simulation under a realistic load profile. Finally, detailed comparison results show that the two-level energy management strategy can achieve lower hydrogen consumption, compared with the rule-based method.

Wireless power transfer (WPT) has shown its potential over conventional charging systems in recent years. However, it is still challenging to determine the power distribution of a multiple-receiver WPT system for its high sensitivity, complex coupling and load relationships. This paper discusses a game theory based control approach for the power distribution of a multiple-receiver WPT system. The power receiver systems (i.e., a receiving coil, a DC-AC rectifier, a DC-DC converter, and an ultracapacitor pack in this paper.) are modelled as independent agents with different preferences under the Matlab simulation environment where the preferences of each agent are represented by utility functions. Then a non-cooperative power distribution game is set up and a generalized Nash equilibrium is found which is used as the reference solution formula to be updated at every control instant. Meanwhile, the generalized Nash equilibrium is found through finding out the variational equilibrium by Karush Kuhn-Tucker conditions (KKT) conditions. The simulation results show that the game theory based control is comparable to the highest efficiency impedance distribution approach in a three-receiver WPT system.

This paper provides an equivalent series resistance-based real-time control method for the battery-ultracapacitor hybrid system. The idea of this control method is that the dynamic load demand is distributed based on the equivalent series resistance ratio of batteries to ultracapacitors, whilst the estimated average load demand based on the past N seconds is supplied by the batteries. In addition, the energy stored in the ultracapacitors is considered for protection purpose. The simulation results verify the effectiveness of the equivalent series resistance-based real-time control method, in terms of the system efficiency, the overall energy loss, and the utilization of the ultracapacitors. Further comparison results show that the efficiency of the proposed real-time control method with well-selected parameter is only 1% lower than that of the dynamic programming method.

This paper discusses a utility function-based control of a battery-ultracapacitor (UC) hybrid energy system. The example system employs the battery semiactive topology. In order to represent different performance and requirements of the battery and UC packs, the two packs are modeled as two independent but related agents using the NetLogo environment. Utility functions are designed to describe the respective preferences of battery and UC packs. Then, the control problem is converted to a multiobjective optimization problem solved by using the Karush-Kuhn-Tucker (KKT) conditions. The weights in the objective functions are chosen based on the location of the knee point in the Pareto set. Both the simulation and experimental results show the utility function-based control provides a comparable performance with the ideal average load demand (ALD)-based control, while the exact preknowledge of the future load demand is not required. The utility function-based control is fast enough to be directly implemented in real time. The discussion in this paper gives a starting point and initial results for dealing with more complex hybrid energy systems.

This paper discusses a game theory based control of an engine/generator-battery-ultracapacitor hybrid energy system. The engine/generator, the battery and ultracapacitor packs are modelled as three independent agents to show their different preferences under the Netlogo environment. The preferences of these three energy components are represented by utility functions. Then a non-cooperative current control game is set up and a Nash equilibrium is found which is used as reference solution formula to be updated at each control instant. The weights in the utility functions are chosen based on the location of the knee point in the Pareto set. The simulation results show that the game theory based control has a comparable performance with the average load demand-based control without knowing the pre-knowledge of the test trip.

This paper provides a quantitative and comparative study on efficiencies of a battery semi-active hybrid energy storage system (HESS) and a battery-only system. The discussion is based on the equivalent series resistance (ESR) circuit models and a pulsed current load. It is theoretically proved that the efficiency difference between the two systems depends on the variance of the load current, the average load current, and the internal resistance of the battery pack. A threshold variance is then accurately derived, above which the battery semi-active HESS becomes more energy efficient. This threshold decreases with an increasing internal resistance of the battery pack and/or a decreasing average load current. Therefore, the best usage of ultracapacitors is to combine with batteries designed for low cost (i.e. a large internal resistance), and supply the dynamic part of a current load with a high peak-to-average ratio. Finally, JC08 driving cycle-based simulation validates the theoretical analysis.

Batteries have been widely used as electrical energy storage units nowadays. However, due to their low power-density, it is usually necessary to combine batteries with other energy storage units, such as super-capacitors, in hybrid energy systems. In this paper, an optimization based control strategy is proposed to improve the energy efficiency as well as battery life time for battery semi-active hybrid systems. Sharing the similar idea as average current strategy but without any predefined driving cycle, this strategy aims to converge the current of the battery pack to the average current of the test trip by mainly enforcing the energy left in the super-capacitor pack to be same as its initial state while to maintain the current variation as small as possible. To achieve those two objectives, a two-objective optimization problem, whose objectives represents the preferences of the battery and supercapacitor packs, is formed and easily solved by using KKT conditions at any control point. The simulation results of this strategy are compared with those from average current strategy, and show that the proposed strategy can achieve comparable performance.