Optimal vehicle dynamics and powertrain control of carbon-free autonomous vehicles: Large language model assisted heterogeneous-agent learning

A novel steering controller for autonomous vehicles is proposed in this paper. For autonomous vehicles, the efficiency is an important indicator, in order to reduce the training cost and improve the real-time performance while the car is running on the road, Offline plus online training is proposed in this paper. Offline plus online training is presented to build a model network of vehicle steering and correct it in real time. Then we adopt a policy iteration adaptive dynamic programming (ADP) to get the optimal control law and deploy it to autonomous vehicles. Eventually, the performance of the controller was evaluated using Matlab simulation. Results verify the feasibility of the theory and show the steering controller is one kind of effective, smooth controller for autonomous vehicles' system.

Navigation control for electric vehicles using nonlinear state feedback [formula omitted] control

This paper focuses on Cooperative Adaptive Cruise Control (CACC) in autonomous vehicles. In CACC, vehicles regulate their speed according to a preceding “leader” vehicle in the lane, forming a platoon. In a benign environment, CACC reduces fuel consumption, maximizes road capacity, and ensures traffic safety. However, CACC is vulnerable to various security threats. In this paper, we consider one of the critical threats, where the platoon leader is compromised, and forges acceleration information sent to platoon members. Such attack would lead to traffic instability and potential collisions. First, we propose information sharing in CACC model to allow vehicles and fixed infrastructure to sense and share information about platoon leaders, hence improves the reliability and supports the detection of anomalous behavior. Then, we propose a real-time anomaly detection mechanism that combines statistical learning with the physics laws of kinematics. Specifically, we propose Generalized Extreme Studentized Deviate with Sliding Chunks (GESD-SC) approach, which is applied at each vehicle in the platoon to detect anomalies in real-time based on the vehicle's own speeding decisions. Kinematic model is also utilized to detect unexpected deviations using the leader's information, communicated directly and observed by the leader's neighboring vehicle(s) and/or supporting infrastructure. Combining kinematic model with GESD-SC has shown to be effective in detecting falsification attacks in CACC. Furthermore, we analyze the time performance, and show that the proposed technique outperforms existing method in detection accuracy and processing time.

The problem of navigation and control for autonomous vehicles is considered. Mutual interactions among vehicle motion dynamics are evaluated. It is proposed the mathematical model suitable for describing and simulating the whole motion of autonomous passenger vehicles.

The problem of navigation and control for autonomous vehicles is considered. Mutual interactions among vehicle motion dynamics are evaluated. It is proposed the mathematical model suitable for describing and simulating the whole motion of autonomous passenger vehicles.

This paper investigates the lane keeping control of autonomous ground vehicles (AGVs) considering the rollover prevention and input saturation. An enhanced state observer-based sliding mode control (SMC) strategy is proposed to achieve the control purpose and maintain the lane keeping errors as well as the roll angle within the prescribed performance boundaries. Three contributions are made in this paper. First, a prescribed performance function (PPF) is proposed in the controller design, aiming to implement the error transformation so as to constrain the controlled variables within the prescribed performance boundaries. Second, a modified sliding surface is developed incorporating two nonlinear functions, whose specialities and benefits are taken advantage of: one is a barrier function to restrict the load transfer ratio (LTR) in a safe boundary to guarantee the roll stability; another is a monotonely decreasing function to adaptively change the damping ratio of the closed-loop system to improve the transient performance, including reducing the transient overshoots and steady-state errors. Third, a modified multivariable adaptive SMC controller is proposed to achieve the integrated lane-keeping and roll control in the presence of the input saturation and bound-unknown disturbances. The stability of the closed-loop system is rigorously proved via the Lyapunov function. Finally, the effectiveness of the proposed control strategy is verified with a high-fidelity and full-car model via the CarSim platform.

: Motion planning and control for autonomous vehicles are complex tasks that must be done in order for a ground robot to operate in a cluttered environment. This dissertation presents the theory, implementation, and test results for some new and novel Receding Horizon Control (RHC) techniques that allow these tasks to be unified into one. The first new method is called Heuristic Receding Horizon Control (HRHC), and uses a modified A* search to fulfill the online optimization required by RHC. The second is called Dual-Frequency Receding Horizon Control (DFRHC), and is used to simplify the trajectory planning process during the RHC optimization. Both methods are combined together to form a practical implementation, which is discussed in detail. The autonomous ground vehicle, the NaviGator, developed at the Center for Intelligent Machines and Robotics, serves as a platform for the implementation and testing discussed.

Technology advancement for on-road vehicles has gained significant momentum in the past decades, particularly in the field of vehicle automation and powertrain electrification. The optimization of powertrain controls for autonomous vehicles typically involves a separated consideration of the vehicle’s external dynamics and powertrain dynamics, with one key aspect often overlooked. This aspect, known as flexible power demand, recognizes that the powertrain control system does not necessarily have to precisely match the power requested by the vehicle motion controller at all times. Leveraging this feature can lead to control designs achieving improved fuel economy by adding an extra degrees-of-freedom to the powertrain control while maintaining safety and drive comfort. The present research investigates the use of an approximate dynamic programming (ADP) approach to develop a powertrain controller, which takes into account the flexibility in power demand within the ADP framework. The concept of reachable sets is incorporated into the ADP framework to ensure safety, improve ride comfort, and enhance the accuracy of the optimization solution. The formulation is based on an autonomous hybrid electric vehicle, while the methodology can also be applied to other types of vehicles. It is also found that necessary customization of the ADP algorithm is needed for this particular control problem to prevent convergence issues. Finally, a case study is presented to evaluate the effectiveness of flexible power demand, as addressed by the ADP method. The experiment demonstrates a 14.1% improvement in fuel economy compared to a scenario without flexible power demand.

A survey on autonomous vehicle control in the era of mixed-autonomy: From physics-based to AI-guided driving policy learning

This paper deals with adaptive control of a nonlinear autonomous underwater vehicle (AUV) in the presence of actuator saturation. Despite the importance of actuator saturation as a practical involvement in the control of autonomous vehicles, it has been considered less in control of AUVs. Therefore, the adaptive $$\mu $$ -modification control method is utilized in this paper for pitch channel autopilot of the REMUS AUV. The designed adaptive $$\mu $$ -modified controller has been applied to the nonlinear six degrees of freedom model of the vehicle. Coefficients of the model are assumed unknown to be coped with an adaptive controller. Performance of the designed controller is compared with that of a direct Model reference adaptive control (MRAC) in six degrees of freedom through simulations. Problems of the MRAC in the presence of input constraints are shown. Shortcomings of conventional adaptive methods in the presence of input constraints are studied. It is finally deduced that how the adaptive controller with $$\mu $$ -modification can perform suitable performance in the presence of the actuator saturation.

Autonomous Vehicle Control based on HoloLens Technology and Raspberry Pi Platform: an Educational Perspective

The widely-used algorithms for motion planning and control of autonomous vehicles usually adopt a layered structure. However, consistency between the two levels is critical to obtain high control performance and maneuvering ability. Aiming at this issue, we propose a two-layer model predictive control (MPC) approach for consistent planning and control of autonomous vehicles. In the higher layer, we propose a finite-horizon convex optimization algorithm for time-optimal motion planning using obstacle convexification of non-convex obstacles. In the proposed planner, the vehicle model uncertainty was learned by a radial-basis-function-based (RBF) neural network whose weight is optimized using pre-collected motion datasets. The surrounding obstacles were detected by a radar, and a clustering algorithm was utilized to process the radar point clouds to isolate polygonal obstacles. Also, the obstacle avoidance constraints were convexified with polyhedrons. As such, the planning problem at the higher layer is transformed into a convex optimization problem to be solved. In the lower layer, a model predictive controller is designed to follow the planned trajectory using the learned dynamic model and a similar performance index to the higher layer. Consequently, the inconsistency between the higher layer and the lower layer can be highly reduced and the control performance can be improved compared with the planning scenario with a nominal model. Extensive simulated and experimental results on planning tasks of unstructured environments have been performed to verify the effectiveness of the proposed approach.

<div class="section abstract"><div class="htmlview paragraph">Modern electrified vehicles rely on drivers to manually adjust control parameters to modify the vehicle's powertrain, such as regenerative braking strength selection or drive mode selection. However, this reliance on infrequent driver input may lead to a mismatch between the selected powertrain control modifiers and the true driving environment. It is therefore advantageous for an electric vehicle's powertrain controller to make online identifications of the current driving conditions. This paper proposes an online driving condition identification scheme that labels drive cycle intervals collected in real-time based on a clustering model, with the objective of informing adaptive powertrain control strategies. HDBSCAN and K-means clustering models are fitted to a data set of drive cycle intervals representing a full range of characteristic driving conditions. The cluster centroids are recorded and used in a vehicle controller to assign driving condition identification labels to the most recently recorded interval of vehicle data. The accuracy of the driving condition identifications of each model is compared by deploying the online identification scheme on the powertrain controller of an electrified vehicle and performing a real-world drive cycle of known driving conditions. The HDBSCAN clusters resulted in superior online driving condition identifications compared to alternative schemes. The main contribution of this paper is the novel application of clustering in an online identification scheme for use in a real-world embedded vehicle controller. By enabling accurate online identification of driving conditions, this approach can improve the powertrain control strategies of electrified vehicles and enhance the driving experience. Future research can leverage the online identification of driving conditions and explore the use of subsequent adaptive control schemes for reducing energy consumption, enhancing safety, and advancing the development of intelligent transportation systems.</div></div>

In the recent years there has been significant effort in the design of intelligent autonomous vehicles capable of operating in variable conditions. The precise modeling of the vehicles dynamics improves the efficiency of vehicles controllers in adverse cases, for example in high velocity, when performing abrupt maneuvers, under mass and loads changes or when moving on rough terrain. Using model-based control approaches it is possible to design a nonlinear controller that maintains the vehicle’s motion characteristics according to given specifications. When the vehicle’s dynamics is subject to modeling uncertainties or when there are unknown forces and torques exerted on the vehicle it is important to be in position to estimate in real-time disturbances and unknown dynamics so as to compensate for them. In this direction, estimation for the unknown dynamics of the vehicle and state estimation-based control schemes have been developed. Feedback control of robotic ground vehicles can be primarily based on (i) global linearization approaches, (ii) approximate linearization approaches and (iii) Lyapunov methods. The control is applied to (i) 4-wheel vehicles models, and (ii) articulated vehicles. At a second stage, to implement control under model uncertainty, estimation methods can be employed capable of identifying in real-time the vehicles’ dynamics. The outcome of the estimation procedure can be used by the aforementioned feedback controllers thus implementing indirect adaptive control schemes. Finally to implement control of the ground vehicles through the measurement of a small number of its state variables, elaborated nonlinear filtering approaches are developed. The topics treated by the chapter are: (a) Nonlinear optimal control of four-wheel autonomous ground vehicles (b) Nonlinear optimal control for an autonomous truck and trailer system (c) Nonlinear optimal control of four-wheel steering autonomous vehicles and (d) Flatness-based control of autonomous four-wheel ground vehicles.

The chapter analyzes the use of nonlinear filtering and control methods based on differential flatness theory in steering control, localization and autonomous navigation of land vehicles, unmanned surface vessels and unmanned aerial vehicles. It is shown that through the application of differential flatness theory one can obtain solution for the following non-trivial problems: state estimation-based control of autonomous vehicles, state estimation-based control of cooperating vehicles, distributed fault diagnosis for autonomous vehicles, velocity control of 4-wheel autonomous vehicles under model uncertainties and external disturbances, active control of vehicle suspensions, state estimation-based control unmanned aerial vehicles of the quadrotor type, state estimation-based control of unmanned surface vessels of the hovercraft type.