One of the main technological stumbling blocks in the field of environmentally friendly vehicles is related to the energy storage system. It is in this regard that car manufacturers are mobilizing to improve battery technologies and to accurately predict their behavior. The work proposed in this paper deals with the advanced electro-thermal modeling of a hybrid energy storage system integrating Lithium-ion batteries and supercapacitors. The objective is to allow the aging aspects of the components of this system to be taken into account. The development of a model including the electro-thermal behaviors makes it possible to evaluate the progressive degradation of the performance of the hybrid energy storage system. The characterization of both components constituting the hybrid system is carried out via a hybrid Particle Swarm–Nelder–Mead (PSO–NM) optimization algorithm using the experimental data of an urban electric vehicle. The obtained results show the good performance of the developed model and confirm the feasibility of our approach. The use of the PSO–NM optimization algorithm facilitated the identification of the parameters of the developed model with high efficiency, as the error observed is less than 3%. The advanced model associated with an adapted sizing method can be used in many cases to compare energy management strategies in electric vehicle applications.

In this paper, a dynamic model of Li-ion batteries incorporating electrothermal and ageing aspects is proposed for electric vehicle applications. The main goal of the proposed model is to be both simple and sufficiently representative of the physical phenomena occurring in a battery cell. These two features allow for using this model as an evaluation tool of electric vehicle performances under different operational and environmental conditions. The developed model is based on an equivalent circuit diagram coupled with a thermal circuit and a semi-empirical ageing equation. Identification of parameters in the dynamic model is conducted by measurement tests in time domain, which uses a hybrid particle Swarm-Nelder-Mead optimization algorithm to achieve excellent prediction over the whole applicable current and state of charge ranges. The validation results show that the proposed model is able to simulate the dynamic interaction between the battery ageing and the thermal as well as electric behavior with sufficient accuracy in the range tested.

Combining a high power source like a supercapacitor with a Li-ion battery for electric vehicle applications results in good performance improvements, highly efficient, long lifetime, lightweight design and relatively modest cost of the overall source. A hybrid energy storage system controlled by a smart energy management strategy can play a key role in the design and development of multi-source electric vehicles. In this work, an optimal energy management strategy based on particle swarm optimization incorporating Nelder-Mead simplex method is proposed. The goal of the proposed strategy is to minimize the battery power stress and improves its lifetime. This is achieved by coupling a rule-based method based on the knowledge of the battery and supercapacitor efficiency operating with a hybrid Particle Swarm–Nelder–Mead (PSO–NM) optimization algorithm. This latter approach is proposed to optimize the control parameters of the rule-based energy management strategy, once the off-line optimization algorithm is over, the control method can be implemented on-line. The obtained results demonstrate significant lifetime enhancements for Li-Ion battery, an increase of up to 20% as compared to the mono-source based on regular single battery.

A light-gray-box simulation model is proposed for a multi-rotor propulsion chain designed to inject various faults into the system to generate fault behavior data sets. The model integrates the electrical and thermal characteristics of each component to model the interactions between components when faults are introduced. The simulated model is compared with an experimental test bench incorporating real hardware components, on which various faults can be injected to assess the accuracy of the simulated model. The propulsion chain consists of an Electronic Speed Controller, a Permanent Magnet Synchronous Motor, and a propeller. The faults considered are introduced on the three components of the propulsion chain. Electronic Speed Controller faults are applied to the power transistors; motor failures are open circuit, overheating and friction; and propeller failures are blade breakage and propeller loss. Using simulation of components with various failures, the model is designed to generate scalable, systematic, and reproducible behavior datasets for the development of fault detection systems which would eliminate the need for real hardware components.

As we advance in the fast-growing era of Machine Learning, various new and more complex neural architectures are arising to tackle problem more efficiently. On the one hand, efficiently deploy them requires advanced knowledge and expertise, which is most of the time difficult to find on the labor market. On the other hand, searching for an optimized neural architecture is a time-consuming task when it is performed manually using a trial-and-error approach. Hence, a method and a tool support are needed to assist users of neural architectures, leading to an eagerness in the field of Automatic Machine Learning (AutoML). When it comes to Deep Learning, an important part of AutoML is the Neural Architecture Search (NAS). In this paper, we propose a formalization for a cell-based search space. The objectives of the proposed approach are to optimize the search-time and to be general enough to handle most of state-of-the-art Convolutional Neural Networks (CNN) architectures, as well as being customizable.

This work proposes an innovative approach to enhance the localization of unmanned aerial vehicles (UAVs) in dynamic environments. The methodology integrates a sophisticated object-tracking algorithm to augment the established simultaneous localization and mapping (ORB-SLAM) framework, utilizing only a monocular camera setup. Moving objects are detected by harnessing the power of YOLOv4, and a specialized Kalman filter is employed for tracking. The algorithm is integrated into the ORB-SLAM framework to improve UAV pose estimation by correcting the impact of moving elements and effectively removing features connected to dynamic elements from the ORB-SLAM process. Finally, the results obtained are recorded using the TUM RGB-D dataset. The results demonstrate that the proposed algorithm can effectively enhance the accuracy of pose estimation and exhibits high accuracy and robustness in real dynamic scenes.