Pouria Sarhadi

Renewed interest in supersonic air travel has prompted researchers to reconsider the design and operation of supersonic transport aircraft. Previously, such aircraft were restricted to overwater routes due to the disturbances caused by their sonic booms. Now, however, low-boom designs and overland flight at marginally supersonic Mach numbers are seen as potential enablers for widespread supersonic air travel. As a result, the trajectories that next-generation supersonic transports may fly are likely to be less constrained than for previous types, and in the last decade there has been a noticeable increase in research focusing on trajectory planning for such aircraft. This paper reviews the different methods that have been used to generate and optimise the flight paths of past and future supersonic transports. The challenges associated with optimising trajectories for aircraft that do not yet exist are discussed, and suggestions for future research activity are presented. Climate-optimal trajectory planning and development of detailed, non-proprietary supersonic aircraft performance models are identified as two key areas for future work.

Renewed interest in supersonic air travel has prompted researchers to consider where next-generation civil supersonic aircraft will operate. These aircraft are expected to fly lower and slower than Concorde, meaning that prevailing winds will have a greater impact on flight duration and fuel consumption. However, nonstandard atmospheric conditions have not always been considered in recent supersonic design studies. This paper therefore quantifies the influence of the jet stream on next-generation supersonic operations between London and New York and investigates the potential benefits of wind-optimal transatlantic routing. Four notional supersonic aircraft with performance characteristics representative of concepts found in the literature are used to assess the effect of wind at different cruise Mach numbers and altitudes. For great circle routes, the assumption of standard atmospheric conditions leads to average time and fuel burn underestimates of 4–7% and 3–6%, respectively, for westbound winter flights. If aircraft deviate from the great circle to seek favorable wind conditions, statistically significant reductions in average flight time and fuel burn are possible, but this is not the case for all cruise altitudes, Mach numbers, or flight directions. Nevertheless, wind-optimal routing is presented as one strategy to help reduce the environmental impact of future supersonic aircraft.

Torque vectoring can enhance dynamic stability and concurrently enable efficient energy management in electric vehicles (EVs) through optimized torque distribution. Nevertheless, conventional torque vectoring schemes often rely on fixed models and tuning, limiting their adaptability. Reinforcement learning (RL) and its model-free versions employing deep neural networks allow the development of control policies through direct interaction with the environment, making it suitable for complex and nonlinear dynamics. This paper presents a comprehensive survey of recent research on the application of RL for torque vectoring and energy optimization in EVs. An overview of conventional direct yaw control (DYC) approaches, their objectives, and common hierarchical strategies are initially studied to establish a foundation for discussing model-free RL-based torque vectoring. A description of RL in the context of stability-oriented control and energy optimization, key components, operational processes, and their classifications are studied. The primary emphasis is on RL-based torque vectoring and energy management in EVs to improve yaw stability, reduce energy consumption, and manage trade-offs under real-time constraints. Overall, RL-based controllers provide enhanced adaptability to modeling inaccuracies and facilitate more straightforward multi-objective design for simultaneous energy management and stability control, making them promising alternatives to conventional model-based methods.

This study presents an innovative solution for simultaneous energy optimization and dynamic yaw control of all-wheel-drive (AWD) electric vehicles (EVs) using deep reinforcement learning (DRL) techniques. To this end, three model-free DRL-based methods, based on deep deterministic policy gradient (DDPG), twin delayed deep deterministic policy gradient (TD3), and TD3 enhanced with curriculum learning (CL TD3), are developed for determining optimal yaw moment control and energy optimization online. The proposed DRL controllers are benchmarked against model-based controllers, i.e., linear quadratic regulator with the sequential quadratic programming (LSQP) and sliding mode control with SQP (SSQP). A tailored multi-term reward function is structured to penalize excessive yaw rate error, sideslip angle, tire slip deviations beyond peak grip regions, and power losses based on a realistic electric machine efficiency map. The learning environment is based on a nonlinear double-track vehicle model, incorporating tire-road interactions. To evaluate the generalizability of the algorithms, the agents are tested across various velocities, tire-road friction coefficients, and additional scenarios implemented in IPG CarMaker, a high-fidelity vehicle dynamics simulator. In addition to the deployment without requiring an explicit model of the plant, the simulation results demonstrate that the proposed solution modifies vehicle dynamics and maneuverability in most cases compared to the model-based conventional controller. Furthermore, the reduction in sideslip angle, excellent traction through minimizing tire slip ratio, avoiding oversteering and understeering, and maintaining an acceptable range of energy optimization are demonstrated for DRL controllers, especially for the TD3 and CL TD3 algorithms.

Over the past decade, remarkable progress has been made in adopting deep neural networks to enhance the performance of conventional reinforcement learning. A notable milestone was the development of Deep Q-Networks (DQN), which achieved human-level performance across a range of Atari games, demonstrating the potential of deep learning to stabilise and scale reinforcement learning. Subsequently, extensions to continuous control algorithms paved the way for a new paradigm in control, one that has attracted broader attention than any classical control approach in recent literature. These developments also demonstrated strong potential for advancing data-driven, model-free algorithms for control and for achieving higher levels of autonomy. However, the application of these methods has remained largely confined to simulated and gaming environments, with ongoing efforts to extend them to real-world applications. Before such deployment can be realised, a solid and quantitative understanding of their performance on applied control problems is necessary. This paper conducts a comparative analysis of these approaches on four diverse benchmark problems with implementation results. This analysis offers a scrutinising and systematic evaluation to shed light on the real-world capabilities and limitations of deep reinforcement learning methods in applied control settings.

A novel artificial intelligence-based approach for the direct yaw control (DYC) of an all-wheel drive (AWD) electric vehicle (EV) is proposed in this paper. To improve adaptability and ability to handle nonlinearities via continuous learning, the proposed algorithm is built upon a twin delayed deep deterministic policy gradient (TD3) reinforcement learning (RL) algorithm for the optimal torque distribution across four wheels of the vehicle. The proposed model-free torque vectoring algorithm performs based on the interaction of an agent with an environment to learn the optimal policy in a reward-driven manner and obtain the ability to dynamically adapt to varying conditions, such as different roads and vehicle speeds. Unlike conventional control methods that rely on precise system modeling and may struggle to adapt under varying conditions, no model of the vehicle is required in the proposed method. This work proposes a model-free RL-based controller with curriculum learning to train the strategy, where the model learns simpler tasks first, progressively increasing difficulty to enhance stability and convergence. A detailed reward function and well-structured actor-critic networks are devised, and the proposed algorithm is compared with a conventional model-based linear quadratic regulator (LQR) approach. A nonlinear model with 7 degrees of freedom is used to model the dynamic behavior of the vehicle in MATLAB/Simulink, and the results are further verified through the implementation of IPG CarMaker under realistic driving scenarios. The performance of the proposed algorithm is studied across different maneuvers, demonstrating reduced yaw rate error and sideslip angle, resulting in enhanced dynamic stability.

This study explores recent advancements in emerging technologies for Demand Responsive Transport (DRT), focusing on the potential integration of autonomous bus services. We review the historical development and current state of research, extending the discussion into the era of Connected and Automated Mobility (CAM), particularly in mass transit. Key components of modern DRT systems are categorised from multiple perspectives, including vehicle types, infrastructure, devices, simulation approaches, decision-making algorithms, and optimisation factors. Each element is discussed in detail, supported by comparative tables to help readers interpret the results effectively. Our findings highlight the growing interest and ongoing research in this promising domain. We identify common practices across studies and areas needing improvement. To support this, we propose a System-of-Systems (SoS) approach to evaluate the technological maturity of DRT solutions, in line with our ongoing project. A System Readiness Level (SRL) analysis is performed by identifying constituent systems and assessing their Technology Readiness Levels (TRLs) and Integration Readiness Levels (IRLs). The study also explores the deployment of DRT in a specific use case: the Maylands Business Park in Hertfordshire, UK. It aims to serve as a comprehensive reference for addressing various dimensions of DRT in mass transit, particularly in the context of automation. It offers insights into current progress and outlines opportunities for deploying DRT as a cost-effective, scalable solution to improve sustainability in Intelligent Transportation Systems (ITS).