Future automotive powertrains become increasingly complex to achieve optimal and robust performance in real-world conditions. This is accompanied by rapidly escalating time and cost demands for calibrating powertrain control systems using existing map- and model-based methods. This challenge highlights the urgent need for self learning control strategies, which autonomously learn optimal control settings on the road. This work demonstrates the potential of Reinforcement Learning-based self learning control for a battery electric vehicle thermal system with safety constraints. To realize Reinforcement Learning on the tested vehicle, a novel exploration method is implemented, which explicitly deals with system safety and minimizes experiment time. By combining an online-learned Gaussian Progress Regression model and a reciprocal Control Barrier Function, the optimal direction and step size for information-rich actions is determined during exploration. Validated on a vehicle test-bench, the proposed method calibrates the reference generator to optimize steady-state operation of the heat pump system. Safe and robust performance is demonstrated for varying ambient temperature and humidity, achieving a relative error in heat pump efficiency within 2% of the true optimum across all validation points. Compared to conventional map-based control, the Reinforcement Learning-based approach reduces calibration time by 69%.

Reinforcement Learning has achieved huge success with various applications in controlled environments. However, limited application is seen in real-world applications due to challenges in guaranteeing safe system operation, required experiment time, and required a-priori system knowledge and models in existing methods. In this work, we propose a novel exploration method, which addresses simultaneously the challenges associated with safe and time-efficient exploration while dealing with system uncertainty. This method integrates a reciprocal Control Barrier Function and an on-line learned Gaussian Process Regression model. For safe system operation, we leverage the information from the reciprocal Control Barrier Function to limit the step size of the agent’s actions, when approaching the safety boundary. To make this exploration process time-efficient, we use the information gain metrics that are calculated using the estimation of the action-values by an on-line learned Gaussian Process Regression model to determine the direction of the agent’s actions. We demonstrate the potential of our exploration method in simulation and on a vehicle test-bench for efficiency-optimal calibration of a thermal management system for battery electric vehicles. To quantify the benefits in terms of safety, optimality, and time efficiency, we benchmark our exploration method with random and uncertainty-driven exploration methods in a simulation environment. For the studied test case, the proposed exploration method satisfies the safety constraint and it converges to within 1.25% of the true optimal action while requiring 28% and 18% lower experiment time compared to the random and uncertainty-driven exploration methods, respectively. For the proposed method, its performance is also demonstrated on a vehicle test bench. Experimental results show that the maximal thermal system efficiency is realized within 2% of the true optimum, while effectively dealing with the safety constraints.

To increase system robustness and autonomy, in this article, we propose a nonlinear fault estimation filter for a class of linear dynamical systems, subject to structured uncertainty, measurement noise, and system delays, in the presence of additive and multiplicative faults. The proposed filter architecture combines tools from model-based control approaches, regression techniques, and convex optimization. The proposed method estimates the additive and multiplicative faults using a linear residual generator combined with nonlinear regression. An offline simulator allows us to numerically characterize the mismatch between an assumed linear model and a range of alternative linear models that exhibit different levels of structured uncertainty. Moreover, we show how the performance bounds of the estimator, valid in the absence of uncertainty, can be used to determine appropriate countermeasures for measurement noise. In the scope of this work, we focus particularly on a fault estimation problem for Society of Automotive Engineers (SAEs) level 4 automated vehicles, which must remain operational in various cases and cannot rely on the driver. The proposed approach is demonstrated in simulations and in an experimental setting, where it is shown that additive and multiplicative faults can be estimated in a real vehicle under the influence of model uncertainty, measurement noise, and delay.

The transition from advanced driver assistance systems to highly automated vehicles proves to be difficult, as the driver is no longer a safety fallback for the latter systems. One of the main challenges is formed by edge cases in the encountered driving scenarios that trigger functional insufficiencies in automated driving (AD) systems. Functional insufficiencies, for the sake of understanding, may be viewed as an inappropriate understanding of or response to a scenario in an AD system, which in turn causes dangerous vehicle behavior. Prior research suggests that using an architecture capable of including redundant heterogeneous AD systems as separate channels, such as the Safety Shell, can mitigate some of these functional insufficiencies. However, this benefit has only been evaluated in limited and deterministic simulation environments. To overcome this, our objectives in this paper are to (i) develop an experimental method for extensive testing of such architectures, and (ii) to assess the suitability of the Safety Shell architecture to handle edge cases with this new method. Using the developed experimental setup we observe a significant safety and availability increase of the Safety Shell compared to the included individual AD channels in the tested scenarios. Finally, our study provides insight into the requirements for the evaluated AD channels.

Automated Driving (AD) systems have the potential to increase safety, comfort and energy efficiency. Recently, major automotive companies have started testing and validating AD systems (ADS) on public roads. Nevertheless, the commercial deployment and wide adoption of ADS have been moderate, partially due to system functional insufficiencies (FI) that undermine passenger safety and lead to hazardous situations on the road. FIs are defined in ISO 21448 Safety Of The Intended Functionality (SOTIF). FIs are insufficiencies in sensors, actuators and algorithm implementations, including neural networks and probabilistic calculations. Examples of FIs in ADS include inaccurate ego-vehicle localization on the road, incorrect prediction of a cyclist maneuver, unreliable detection of a pedestrian, etc. The main goal of our study is to formulate a generic architectural design pattern, which is compatible with existing methods and ADS, to improve FI mitigation and enable faster commercial deployment of ADS. First, we studied the 2021 autonomous vehicles disengagement reports published by the California Department of Motor Vehicles (DMV). The data clearly show that disengagements are five times more often caused by FIs rather than by system faults. We then made a comprehensive list of insufficiencies and their characteristics by analyzing over 10 hours of publicly available road test videos. In particular, we identified insufficiency types in four major categories: world model, motion plan, traffic rule, and operational design domain. The insufficiency characterization helps making the SOTIF analyses of triggering conditions more systematic and comprehensive. Based on our FI characterization, simulation experiments and literature survey, we define a novel generic architectural design pattern Daruma to dynamically select the channel that is least likely to have a FI at the moment.

Predicting the movement of other road users is beneficial for improving automated vehicle (AV) performance. However, the relationship between the time horizon associated with these predictions and AV performance remains unclear. Despite the existence of numerous trajectory prediction algorithms, no studies have been conducted on how varying prediction lengths affect AV safety and other vehicle performance metrics, resulting in undefined horizon requirements for prediction methods. Our study addresses this gap by examining the effects of different prediction horizons on AV performance, focusing on safety, comfort, and efficiency. Through multiple experiments using a state-of-the-art, risk-based predictive trajectory planner, we simulated predictions with horizons up to 20 seconds. Based on our simulations, we propose a framework for specifying the minimum required and optimal prediction horizons based on specific AV performance criteria and application needs. Our results indicate that a horizon of 1.6 seconds is required to prevent collisions with crossing pedestrians, horizons of 7-8 seconds yield the best efficiency, and horizons up to 15 seconds improve passenger comfort. We conclude that prediction horizon requirements are application-dependent, and recommend aiming for a prediction horizon of 11.8 seconds as a general guideline for applications involving crossing pedestrians.

Predicting the behavior of road users accurately is crucial to enable the safe operation of autonomous vehicles in urban or densely populated areas. Therefore, there has been a growing interest in time series motion prediction research, leading to significant advancements in state-of-the-art techniques in recent years. However, the potential of using LiDAR data to capture more detailed local features, such as a person's gaze or posture, remains largely unexplored. To address this, we develop a novel multimodal approach for motion prediction based on the PointNet foundation model architecture, incorporating local LiDAR features. Evaluation on the Waymo Open Dataset shows a performance improvement of 6.20 and compared with the previous state-of-the-art MTR. We open-source the code of our LiMTR model.

To improve on road safety when autonomous vehicles (AVs) are introduced for highway or urban driving, in this article, we design an automated merging algorithm for an AV into a mixed-traffic flow scenario (i.e., traffic including autonomous and manually driven vehicles). In particular, we propose a novel model predictive control (MPC)-based solution to perform a merging procedure from a double lane into a single lane and continue with (adaptive) cruise control (A)CC functionality after the merge in one integrated algorithm. The proposed MPC balances fast progress along the path with comfort, while obeying a state-dependent safety distance and velocity bounds. Recursive feasibility, leading to safety and proper behavior (i.e., rigorously satisfying constraints), is guaranteed by the design of proper terminal sets, extending existing terminal sets in the literature. The resulting MPC problem is a mixed-integer quadratic program (MIQP) problem, which can be solved for global optimality. Through numerical simulations and experimental validation of the algorithm with multibrand cars, we demonstrate desirable behavior and verify the effectiveness of the proposed MPC merging scheme.

To enable highly automated vehicles where the driver is no longer a safety backup, the vehicle must deal with various Functional Insufficiencies (FIs). Thus-far, there is no widely accepted functional architecture that maximizes the availability of autonomy and ensures safety in complex vehicle operational design domains. In this paper, we present a survey of existing methods that strive to prevent or handle FIs. We observe that current design-time methods of preventing FIs lack completeness guarantees. Complementary solutions for on-line handling cannot suitably increase safety without seriously impacting availability of journey continuing autonomous functionality. To fill this gap, we propose the Safety Shell, a scalable multi-channel architecture and arbitration design, built upon preexisting functional safety redundant channel architectures. We compare this novel approach to existing architectures using numerical case studies. The results show that the Safety Shell architecture allows the automated vehicle to be as safe or safer compared to alternatives, while simultaneously improving availability of vehicle autonomy, thereby increasing the possible coverage of on-line functional insufficiency handling.

As vehicle automation advances, motion planning algorithms face escalating challenges in achieving safe and efficient navigation. Existing Advanced Driver Assistance Systems (ADAS) primarily focus on basic tasks, leaving unexpected scenarios for human intervention, which can be error-prone. Motion planning approaches for higher levels of automation in the state-of-the-art are primarily oriented toward the use of risk- or anti-collision constraints, using over-approximates of the shapes and sizes of other road users to prevent collisions. These methods however suffer from conservative behavior and the risk of infeasibility in high-risk initial conditions. In contrast, our work introduces a novel multi-objective trajectory generation approach. We propose an innovative method for constructing risk fields that accommodates diverse entity shapes and sizes, which allows us to also account for the presence of potentially occluded objects. This methodology is integrated into an occlusion-aware trajectory generator, enabling dynamic and safe maneuvering through intricate environments while anticipating (potentially hidden) road users and traveling along the infrastructure toward a specific goal. Through theoretical underpinnings and simulations, we validate the effectiveness of our approach. This paper bridges crucial gaps in motion planning for automated vehicles, offering a pathway toward safer and more adaptable autonomous navigation in complex urban contexts.

Autonomous vehicles require accurate, real-time environmental representations for safe navigation. Traditional methods relying on detailed offline maps struggle in dynamic environments with outdated data. Thus, real-time solutions integrating diverse data sources and adapting to current situations are essential. An existing framework addressing this challenge is the situation-aware drivable space (SDS). However, SDS faces limitations such as simplified object representation, non-standard output, and validation only on simulated or high-quality data. This paper introduces SDS++, an advanced framework for drivable space estimation addressing these limitations. Key improvements include incorporating complex geometries, providing a more flexible and accurate drivable space, and adopting a standardized output representation. Methodologically, SDS++ operates in two main steps. First, it formulates a Graph-based SLAM optimization from semantic objects and determines the most likely environment, leveraging a novel custom factor for optimizing curved line features. Second, it estimates drivable space using artificial potential fields (APFs) constructed from domain knowledge through implicit and sigmoid functions. SDS++ offers several benefits. The novel factor allows the incorporation of complex geometries such as road lanes and edges. Constructing APFs from domain knowledge provides a flexible drivable space that adapts to the driving situation. Additionally, APFs created from implicit and sigmoid functions offer higher accuracy than typical bi-Gaussian representations. SDS++ has been rigorously validated in simulations and through experiments in a real vehicle with unrefined data. Integration with a model predictive control (MPC)-based planner demonstrates significant enhancements in trajectory planning, showcasing SDS++'s robustness against localization noise and ability to adapt trajectories to the driving context.

In order to make the complex driving task of merging safer, in this paper we consider the automated merging of an autonomous vehicle into a mixed-traffic flow scenario (i.e., traffic including autonomous and manually driven vehicles). In particular, we propose a novel MPC-based algorithm to perform a merging procedure from a double lane into a single lane and continue with (adaptive) cruise control ((A)CC) functionality after the merge. The proposed MPC balances fast progress along the path with comfort, while obeying safety and maximum allowed velocity bounds. Recursive feasibility, leading to safety and proper behavior, is guaranteed by the design of a proper terminal set, extending existing ones in the literature. The on-line MPC problem is translated into a mixed integer quadratic program (MIQP) that can be solved for global optimality. Through numerical simulations we demonstrate the behavior and effectiveness of the proposed MPC merging scheme.

With increased developments and interest in cooperative driving and higher levels of automation (SAE level 3+), the need for safety systems that are capable to monitor system health and maintain safe operations in faulty scenarios is increasing. A variety of faults or failures could occur, and there exists a high variety of ways to respond to such events. Once a fault or failure is detected, there is a need to classify its severity and decide on appropriate and safe mitigating actions. To provide a solution to this mitigation challenge, in this paper a functional-safety architecture is proposed and an optimization-based mitigation algorithm is introduced. This algorithm uses nonlinear model predictive control (NMPC) to bring a vehicle, suffering from a severe fault, such as a power steering failure, to a safe-state. The internal model of the NMPC uses the information from the fault detection, isolation and identification to optimize the tracking performance of the controller, showcasing the need of the proposed architecture. Given a string of ACC vehicles, our results demonstrate a variety of tactical decision-making approaches that a fault-affected vehicle could employ to manage any faults. Furthermore, we show the potential for improving the safety of the affected vehicle as well as the effect of these approaches on the duration of the manoeuvre.

For automotive powertrain control systems, the calibration effort is exploding due to growing system complexity and increasingly strict legal requirements for greenhouse gas and real-world pollutant emissions. These powertrain systems are characterized by their highly dynamic operation, so transient performance is key. Currently applied control methods require tuning of an increasing number of look-up tables and of parameters in the applied models. Especially for transient control this state-of-the-art calibration process is unsystematic and requires a large development effort. Also, embedding models in a controller can set challenging requirements to production control hardware. In this work, we assess the potential of Machine Learning to dramatically reduce the calibration effort in transient air path control development. This is not only done for the existing benchmark controller, but also for a new preview controller. In order to efficiently realize preview, a strategy is proposed where the existing reference signal is shifted in time. These reference signals are then modeled as a function of engine torque demand using a Long Short-Term Memory (LSTM) neural network, which can capture the dynamic input–output relationship. A multi-objective optimization problem is defined to systematically select hyperparameters that optimize the trade-off between model accuracy, system performance, calibration effort and computational requirements. This problem is solved using an exhaustive search approach. The control system performance is validated over a transient driving cycle. For the LSTM-based controllers, the proposed calibration approach achieves a significant reduction of 71% in the control calibration effort compared to the benchmark process. The expert effort and turbocharger experiments used in calibrating transient compensation maps in physics-based feedforward controller are replaced by little simulation time and parametrization effort in ML-based controller, which requires significantly less expert effort and system knowledge compared to benchmark process. The best trade-off between multi-objective cost terms is achieved with one layer and 32 cells LSTM neural network for both non-preview and preview control. For non-preview control, a comparable control system performance is achieved with the LSTM-based controller, while 5% reduction in cumulative NOx emissions and similar fuel consumption is achieved with preview controller.

Accurate and robust trajectory predictions of road users are needed to enable safe automated driving. To do this, machine learning models are often used, which can show erratic behavior when presented with previously unseen inputs. In this work, two environment-aware models (MotionCNN and MultiPath++) and two common baselines (Constant Velocity and an LSTM) are benchmarked for robustness against various perturbations that simulate functional insufficiencies observed during model deployment in a vehicle: unavailability of road information, late detections, and noise. Results show significant performance degradation under the presence of these perturbations, with errors increasing up to +1444.8% in commonly used trajectory prediction evaluation metrics. Training the models with similar perturbations effectively reduces performance degradation, with error increases of up to +87.5%. We argue that despite being an effective mitigation strategy, data augmentation through perturbations during training does not guarantee robustness towards unforeseen perturbations, since identification of all possible on-road complications is unfeasible. Furthermore, degrading the inputs sometimes leads to more accurate predictions, suggesting that the models are unable to learn the true relationships between the different elements in the data.

Automating the driving task could improve traffic safety, since 93% of traffic accidents are human failure related. To enable automated driving, one of the challenges is to plan safe and optimal trajectories. Previous research mentions that future road users' predictions are a prerequisite to achieve safe and high-quality motion planning for autonomous driving. This is a possible claim, yet no research exists that explains to what extend the availability prediction is useful. To quantifiable investigate the claim that other road users' predictions indeed result in safer and comfortable traffic, in this paper a safety and comfort analysis of different highway scenarios is performed for different levels of automation. Our simulation-based analysis shows that the availability of other road users' predictions indeed leads to increased safety in different lane merging scenarios (i.e., merging and traffic jam; short and long ramps). Further, the effect of predictions is more visible in higher level automated cars. However, comfort improvements with predictions was not witnessed in the data, due to a balance between comfort and safety (i.e., cases that become safe are not necessarily comfortable). Less comfortable, but safe, faster trajectories are preferred during the simulation with predictions (i.e., trajectories with higher velocity and accelerations).

The urban environment is amongst the most difficult domains for autonomous vehicles. The vehicle must be able to plan a safe route on challenging road layouts, in the presence of various dynamic traffic participants such as vehicles, cyclists and pedestrians and in various environmental conditions. The challenge remains to have motion planners that are computationally fast and that account for future movements of other road users proactively. This paper describes an computationally efficient sampling-based trajectory planner for safe and comfortable driving in urban environments. The planner improves the Stable-Sparse-RRT algorithm by adding initial exploration branches to the search tree based on road layout information and reiterating the previous solution. Furthermore, the trajectory planner accounts for the predicted motion of other traffic participants to allow for safe driving in urban traffic. Simulation studies show that the planner is capable of planning collision-free, comfortable trajectories in several urban traffic scenarios. Adding the domain-knowledge-based exploration branches increases the efficiency of exploration of highly interesting areas, thereby increasing the overall planning performance.

Developing safe automated vehicles that can be proactive, safe, and comfortable in mixed traffic requires improved planning methods that are risk-averse and that account for predictions of the motion of other road users. To consider these criteria, in this article, we propose a non-linear model-predictive trajectory generator scheme, which couples the longitudinal and lateral motion of the vehicle to steer the vehicle with minimal risk, while progressing towards the goal state. The proposed method takes into account the infrastructure, surrounding objects, and predictions of the objects' state through artificial potential-based risk fields included in the cost function of the model-predictive control (MPC) problem. This trajectory generator enables anticipatory maneuvers, i.e., mitigating risk far before any safety-critical intervention would be necessary. The method is proven in several case studies representing both highways- and urban situations. The results show the safe and efficient implementation of the MPC trajectory generator while proving its real-time applicability.

Complexity of engine control systems is continuously growing due to an increased number of subsystems and the need for robust performance. For traditional map-based as well as state-of-the-art model-based approaches, this will lead to unacceptable development costs and time for future engines. Parametrization of the embedded models using supervised learning regression methods can immensely reduce the number of calibration parameters and hence the calibration effort. However, a methodology for performance assessment of different promising data-driven modelling methods for engine control development is currently lacking. In this paper, a systematic methodology that assesses model inaccuracy, and also implementation aspects such as calibration effort and computational complexity is introduced. This method is applied to assess the potential of Supervised Learning (SL) methods for parametrizing the feedforward controller of a modern diesel engine air-path controller. Using requirements analysis and the specified performance criteria, two regression methods were selected: artificial neural networks (ANN) and support vector machines (SVM). After careful data selection and model training, performance is compared with the benchmark controller, which uses a physics-based model. From simulation results, it is shown that a 97% reduction in the number of calibration parameters with both regression models can be realized. For a standard test cycle, cumulative engine out NOx emissions with regression based controllers are close to the allowable inaccuracy of 10% compared to the benchmark controller. Among the two methods, ANN shows the best performance for the studied performance criteria of inaccuracy, number of calibration parameters and computational complexity.

Estimating and detecting faults is crucial in ensuring safe and efficient automated systems. In the presence of disturbances, noise, or varying system dynamics, such estimation is even more challenging. To address this challenge, this letter proposes a novel filter to estimate multiple fault signals for a class of discrete-time linear parameter-varying (LPV) systems. The design of such a filter is formulated as an optimization problem and is solved recursively, while the system dynamics may vary over time. Conditions for the existence and detectability of the fault are introduced and the problem is formulated and solved using the quadratic programming framework. We further propose an approximate scheme that can be arbitrarily precise while it enjoys an analytical solution, which supports real-time implementation. The method is illustrated and validated on an automated vehicle’s lateral dynamics, which is a practically relevant example for LPV systems. The results show that the estimation filter can decouple unknown disturbances and known or measurable parameter variations in the dynamics while estimating the unknown fault.