This paper explores the use of reinforcement learning (RL) models for autonomous racing. In contrast to passenger cars, where safety is the top priority, a racing car aims to minimize the lap-time. We frame the problem as a reinforcement learning task with a multidimensional input consisting of the vehicle telemetry, and a continuous action space. To find out which RL methods better solve the problem and whether the obtained models generalize to driving on unknown tracks, we put 10 variants of deep deterministic policy gradient (DDPG) to race in two experiments: i)~studying how RL methods learn to drive a racing car and ii)~studying how the learning scenario influences the capability of the models to generalize. Our studies show that models trained with RL are not only able to drive faster than the baseline open source handcrafted bots but also generalize to unknown tracks.
We consider autonomous racing of two cars and present an approach to formulate racing decisions as a non-cooperative non-zero-sum game. We design three different games where the players aim to fulfill static track constraints as well as avoid collision with each other; the latter constraint depends on the combined actions of the two players. The difference between the games are the collision constraints and the payoff. In the first game collision avoidance is only considered by the follower, and each player maximizes their own progress towards the finish line. We show that, thanks to the sequential structure of this game, equilibria can be computed through an efficient sequential maximization approach. Further, we show these actions, if feasible, are also a Stackelberg and Nash equilibrium in pure strategies of our second game where both players consider the collision constraints. The payoff of our third game is designed to promote blocking, by additionally rewarding the cars for staying ahead at the end of the horizon. We show that this changes the Stackelberg equilibrium, but has a minor influence on the Nash equilibria. For online implementation, we propose to play the games in a moving horizon fashion, and discuss two methods for guaranteeing feasibility of the resulting coupled repeated games. Finally, we study the performance of the proposed approaches in simulation for a set-up that replicates the miniature race car tested at the Automatic Control Laboratory of ETH Zurich. The simulation study shows that the presented games can successfully model different racing behaviors and generate interesting racing situations.
In this paper we consider autonomous driving of miniature race cars. The viability kernel is used to efficiently generate finite look-ahead trajectories that maximize progress while remaining recursively feasible with respect to static obstacles (e.g., stay inside the track). Together with a low-level model predictive controller, this method makes real-time autonomous racing possible. The viability kernel computation is based on space discretization. To make the calculation robust against discretization errors, we propose a novel numerical scheme based on game theoretical methods, in particular the discriminating kernel. We show that the resulting algorithm provides an inner approximation of the viability kernel and guarantees that, for all states in the cell surrounding a viable grid point, there exists a control that keeps the system within the kernel. The performance of the proposed control method is studied in simulation where we determine the effects of various design choices and parameters and in experiments on an autonomous racing set-up maintained at the Automatic Control Laboratory of ETH Zurich. Both simulation and experimental results suggest that the more conservative approximation using the discriminating kernel results in safer driving style at the cost of a small increase in lap time.
Machine learning and artificial intelligence (ML/AI) are rapidly becoming an indispensable part of physics research, with domain applications ranging from theory and materials prediction to high-throughput data analysis. In parallel, the recent successes in applying ML/AI methods for autonomous systems from robotics through self-driving cars to organic and inorganic synthesis are generating enthusiasm for the potential of these techniques to enable automated and autonomous experiment (AE) in imaging. Here, we aim to analyze the major pathways towards AE in imaging methods with sequential image formation mechanisms, focusing on scanning probe microscopy (SPM) and (scanning) transmission electron microscopy ((S)TEM). We argue that automated experiments should necessarily be discussed in a broader context of the general domain knowledge that both informs the experiment and is increased as the result of the experiment. As such, this analysis should explore the human and ML/AI roles prior to and during the experiment, and consider the latencies, biases, and knowledge priors of the decision-making process. Similarly, such discussion should include the limitations of the existing imaging systems, including intrinsic latencies, non-idealities and drifts comprising both correctable and stochastic components. We further pose that the role of the AE in microscopy is not the exclusion of human operators (as is the case for autonomous driving), but rather automation of routine operations such as microscope tuning, etc., prior to the experiment, and conversion of low latency decision making processes on the time scale spanning from image acquisition to human-level high-order experiment planning.
Reliable navigation like expert human drivers in urban environments is a critical capability for autonomous vehicles. Traditional methods for autonomous driving are implemented with many building blocks from perception, planning and control, making them difficult to generalize to varied scenarios due to complex assumptions and interdependencies. In this paper, we develop an end-to-end trajectory generation method based on imitation learning. It can extract spatiotemporal features from the front-view camera images for scene understanding, then generate collision-free trajectories several seconds into the future. The proposed network consists of three sub-networks, which are selectively activated for three common driving tasks: keep straight, turn left and turn right. The experimental results suggest that under various weather and lighting conditions, our network can reliably generate trajectories in different urban environments, such as turning at intersections and slowing down for collision avoidance. Furthermore, by integrating the proposed network into a navigation system, good generalization performance is presented in an unseen simulated world for autonomous driving on different types of vehicles, such as cars and trucks.
Perception-driven approach and end-to-end system are two major vision-based frameworks for self-driving cars. However, it is difficult to introduce attention and historical information of autonomous driving process, which are the essential factors for achieving human-like driving into these two methods. In this paper, we propose a novel model for self-driving cars named brain-inspired cognitive model with attention (CMA). This model consists of three parts: a convolutional neural network for simulating human visual cortex, a cognitive map built to describe relationships between objects in complex traffic scene and a recurrent neural network that combines with the real-time updated cognitive map to implement attention mechanism and long-short term memory. The benefit of our model is that can accurately solve three tasks simultaneously:1) detection of the free space and boundaries of the current and adjacent lanes. 2)estimation of obstacle distance and vehicle attitude, and 3) learning of driving behavior and decision making from human driver. More significantly, the proposed model could accept external navigating instructions during an end-to-end driving process. For evaluation, we build a large-scale road-vehicle dataset which contains more than forty thousand labeled road images captured by three cameras on our self-driving car. Moreover, human driving activities and vehicle states are recorded in the meanwhile.
Within a decade, almost every major auto company, along with fleet operators such as Uber, have announced plans to put autonomous vehicles on the road. At the same time, electric vehicles are quickly emerging as a next-generation technology that is cost effective, in addition to offering the benefits of reducing the carbon footprint. The combination of a centrally managed fleet of driverless vehicles, along with the operating characteristics of electric vehicles, is creating a transformative new technology that offers significant cost savings with high service levels. This problem involves a dispatch problem for assigning riders to cars, a planning problem for deciding on the fleet size, and a surge pricing problem for deciding on the price per trip. In this work, we propose to use approximate dynamic programming to develop high-quality operational dispatch strategies to determine which car (given the battery level) is best for a particular trip (considering its length and destination), when a car should be recharged, and when it should be re-positioned to a different zone which offers a higher density of trips. We then discuss surge pricing using an adaptive learning approach to decide on the price for each trip. Finally, we discuss the fleet size problem which depends on the previous two problems.
Autonomous driving systems validation remains one of the biggest challenges car manufacturers must tackle in order to provide safe driverless cars. The high complexity stems from several factors: the multiplicity of vehicles, embedded systems, use cases, and the very high required level of reliability for the driving system to be at least as safe as a human driver. In order to circumvent these issues, large scale simulations reproducing this huge variety of physical conditions are intensively used to test driverless cars. Therefore, the validation step produces a massive amount of data, including many time-indexed ones, to be processed. In this context, building a structure in the feature space is mandatory to interpret the various scenarios. In this work, we propose a new co-clustering approach adapted to high-dimensional time series analysis, that extends the standard model-based co-clustering. The FunCLBM model extends the recently proposed Functional Latent Block Model and allows to create a dependency structure between row and column clusters. This structured partition acts as a feature selection method, that provides several clustering views of a dataset, while discriminating irrelevant features. In this workflow, times series are projected onto a common interpolated low-dimensional frequency space, which allows to optimize the projection basis. In addition, FunCLBM refines the definition of each latent block by performing block-wise dimension reduction and feature selection. We propose a SEM-Gibbs algorithm to infer this model, as well as a dedicated criterion to select the optimal nested partition. Experiments on both simulated and real-case Renault datasets shows the effectiveness of the proposed tools and the adequacy to our use case.
Autonomous Vehicles (AVs) i.e., self-driving cars, operate in a safety critical domain, since errors in the autonomous driving software can lead to huge losses. Statistically, road intersections which are a part of the AVs operational design domain (ODD), have some of the highest accident rates. Hence, testing AVs to the limits on road intersections and assuring their safety on road intersections is pertinent, and thus the focus of this paper. We present a situation coverage-based (SitCov) AV-testing framework for the verification and validation (V&V) and safety assurance of AVs, developed in an open-source AV simulator named CARLA. The SitCov AV-testing framework focuses on vehicle-to-vehicle interaction on a road intersection under different environmental and intersection configuration situations, using situation coverage criteria for automatic test suite generation for safety assurance of AVs. We have developed an ontology for intersection situations, and used it to generate a situation hyperspace i.e., the space of all possible situations arising from that ontology. For the evaluation of our SitCov AV-testing framework, we have seeded multiple faults in our ego AV, and compared situation coverage based and random situation generation. We have found that both generation methodologies trigger around the same number of seeded faults, but the situation coverage-based generation tells us a lot more about the weaknesses of the autonomous driving algorithm of our ego AV, especially in edge-cases. Our code is publicly available online, anyone can use our SitCov AV-testing framework and use it or build further on top of it. This paper aims to contribute to the domain of V&V and development of AVs, not only from a theoretical point of view, but also from the viewpoint of an open-source software contribution and releasing a flexible/effective tool for V&V and development of AVs.
Motorsport has always been an enabler for technological advancement, and the same applies to the autonomous driving industry. The team TUM Auton-omous Motorsports will participate in the Indy Autonomous Challenge in Octo-ber 2021 to benchmark its self-driving software-stack by racing one out of ten autonomous Dallara AV-21 racecars at the Indianapolis Motor Speedway. The first part of this paper explains the reasons for entering an autonomous vehicle race from an academic perspective: It allows focusing on several edge cases en-countered by autonomous vehicles, such as challenging evasion maneuvers and unstructured scenarios. At the same time, it is inherently safe due to the motor-sport related track safety precautions. It is therefore an ideal testing ground for the development of autonomous driving algorithms capable of mastering the most challenging and rare situations. In addition, we provide insight into our soft-ware development workflow and present our Hardware-in-the-Loop simulation setup. It is capable of running simulations of up to eight autonomous vehicles in real time. The second part of the paper gives a high-level overview of the soft-ware architecture and covers our development priorities in building a high-per-formance autonomous racing software: maximum sensor detection range, relia-ble handling of multi-vehicle situations, as well as reliable motion control under uncertainty.