In Social Robot Navigation (SRN), the availability of meaningful metrics is crucial for evaluating trajectories from human-robot interactions. In the SRN context, such interactions often relate to resolving conflicts between two or more agents. Correspondingly, the shares to which agents contribute to the resolution of such conflicts are important. This paper builds on recent work, which proposed a Responsibility metric capturing such shares. We extend this framework in two directions: First, we model the conflict buildup phase by introducing a time normalization. Second, we propose the related Engagement metric, which captures how the agents' actions intensify a conflict. In a comprehensive series of simulated scenarios with dyadic, group and crowd interactions, we show that the metrics carry meaningful information about the cooperative resolution of conflicts in interactions. They can be used to assess behavior quality and foresightedness. We extensively discuss applicability, design choices and limitations of the proposed metrics.

In Social Robot Navigation, autonomous agents need to resolve many sequential interactions with other agents. State-of-the art planners can efficiently resolve the next, imminent interaction cooperatively and do not focus on longer planning horizons. This makes it hard to maneuver scenarios where the agent needs to select a good strategy to find gaps or channels in the crowd. We propose to decompose trajectory planning into two separate steps: Conflict avoidance for finding good, macroscopic trajectories, and cooperative collision avoidance (CCA) for resolving the next interaction optimally. We propose the Probabilistic Gap Planner (PGP) as a conflict avoidance planner. PGP modifies an established probabilistic collision risk model to include a general assumption of cooperativity. PGP biases the short-term CCA planner to head towards gaps in the crowd. In extensive simulations with crowds of varying density, we show that using PGP in addition to state-of-the-art CCA planners improves the agents' performance: On average, agents keep more space to others, create less tension, and cause fewer collisions. This typically comes at the expense of slightly longer paths. PGP runs in real-time on WaPOCHI mobile robot by Honda R D.

Establishing standardized metrics for Social Robot Navigation (SRN) algorithms for assessing the quality and social compliance of robot behavior around humans is essential for SRN research. Currently, commonly used evaluation metrics lack the ability to quantify how cooperative an agent behaves in interaction with humans. Concretely, in a simple frontal approach scenario, no metric specifically captures if both agents cooperate or if one agent stays on collision course and the other agent is forced to evade. To address this limitation, we propose two new metrics, a conflict intensity metric and the responsibility metric. Together, these metrics are capable of evaluating the quality of human-robot interactions by showing how much a given algorithm has contributed to reducing a conflict and which agent actually took responsibility of the resolution. This work aims to contribute to the development of a comprehensive and standardized evaluation methodology for SRN, ultimately enhancing the safety, efficiency, and social acceptance of robots in human-centric environments.

Interaction modeling plays a huge role in understanding human behavior in traffic. This is especially relevant when it comes to interactions between vehicles and vulnerable road users such as pedestrians. Thus, pedestrian intention prediction is an ongoing field of research in order to understand the pedestrians' decision making. Most state-of-the-art prediction frameworks are trained on large-scale datasets and evaluated with respect to acknowledged benchmarks. These datasets lack the ability to account for the reciprocal nature of interactions between pedestrians and vehicles and the effects of the two agents influencing each other. In this work, we demonstrate first steps towards assessing pedestrian prediction algorithms within realistic scenarios including the interaction effects arising from its interplay with a planning component. For this, we validate an existing prediction framework trained on benchmark datasets with situations from a virtual reality (VR) pedestrian-vehicle co-simulator that allows us to include the effect of vehicle planning on pedestrian behavior. We evaluate the performance of the prediction framework comparing data from pre-recorded real-world datasets with data from our co-simulation study and conduct an ablation analysis to identify the most important features for pedestrian intention prediction. The results highlight the significance of pedestrian action and proximity to the road.

In interactions between automated vehicles (AVs) and crossing pedestrians, modeling implicit vehicle communication is crucial. In this work, we present a combined prediction and planning approach that allows to consider the influence of the planned vehicle behavior on a pedestrian and predict a pedestrian's reaction. We plan the behavior by solving two consecutive optimal control problems (OCPs) analytically, using variational calculus. We perform a validation step that assesses whether the planned vehicle behavior is adequate to trigger a certain pedestrian reaction, which accounts for the closed-loop characteristics of prediction and planning influencing each other. In this step, we model the influence of the planned vehicle behavior on the pedestrian using a probabilistic behavior acceptance model that returns an estimate for the crossing probability. The probabilistic modeling of the pedestrian reaction facilitates considering the pedestrian's costs, thereby improving cooperative behavior planning. We demonstrate the performance of the proposed approach in simulated vehicle-pedestrian interactions with varying initial settings and highlight the decision making capabilities of the planning approach.

In Automated Driving (AD), traffic scenarios where two agents must resolve an ordering without knowing each other's intention are critical for expanding the operational design domain of automated vehicles to urban environments. These scenarios require negotiation to determine who passes first through an interaction zone. We present a novel agreement measure and negotiation approach to resolve these ordering problems across a wide range of common scenarios. Our method emphasizes detecting and deciding when to switch between potential negotiation outcomes. Our approach extends existing behavior planners to cope with bimodal cooperative interactions, where two potentially desirable outcomes need to be considered. We evaluate our approach by providing both an illustrative scenario and extensive statistical experiments across various geometries, including oncoming narrow passages, crossing and merging scenarios. The results demonstrate that our system considerably improves the behavior in cooperative ordering scenarios compared to the baseline. Furthermore, it is also robust in the sense that it effectively handles dynamic situations where the other agent's intentions changes during the negotiation process.

Interactions between humans and automated vehicles (AVs) will become increasingly common in future traffic. Cooperative behavior enables comfortable and efficient resolutions of such interactions through joint actions. Thus, AVs need the ability to communicate their intention and planned behavior to outside road users to facilitate cooperation among the interaction partners. Many recent studies suggest external human-machine-interfaces (eHMIs) as a feasible way to realize the communication between AVs and other road users. These studies often focus on the design of eHMIs and evaluate their usability based on subjective measures from the perspective of the outside road user. There is only little research on the objective benefits of explicit external communication on the resolution of interactions with AVs. The decision making behind the activation of the communication and the consideration of the actual interaction in the communication are still open research topics. In this work, we present a communication framework that facilitates cooperation between an AV and pedestrians by communicating the vehicle’s yielding intention to give right of way. We demonstrate the proposed system’s ability to improve the joint utility as well as the driving comfort by empirically evaluating simulated interactions between AVs and pedestrians. The results indicate that timely communication leads to more efficient and more predictable pedestrian-vehicle interactions.

Driver support systems that include human states in the support process is an active research field. Many recent approaches allow, for example, to sense the driver's drowsiness or awareness of the driving situation. However, so far, this rich information has not been utilized much for improving the effectiveness of support systems. In this paper, we therefore propose a warning system that uses human states in the form of driver errors and can warn users in some cases of upcoming risks several seconds earlier than the state of the art systems not considering human factors. The system consists of a behavior planner Risk Maps which directly changes its prediction of the surrounding driving situation based on the sensed driver errors. By checking if this driver's behavior plan is objectively safe, a more robust and foresighted driver warning is achieved. In different simulations of a dynamic lane change and intersection scenarios, we show how the driver's behavior plan can become unsafe, given the estimate of driver errors, and experimentally validate the advantages of considering human factors.

Reinforcement Learning (RL) can enable agents to learn complex tasks. However, it is difficult to interpret the knowledge and reuse it across tasks. Inductive biases can address such issues by explicitly providing generic yet useful decomposition that is otherwise difficult or expensive to learn implicitly. For example, object-centered approaches decompose a high dimensional observation into individual objects. Expanding on this, we utilize an inductive bias for explicit object-centered knowledge separation that provides further decomposition into semantic representations and dynamics knowledge. For this, we introduce a semantic module that predicts an objects' semantic state based on its context. The resulting affordance-like object state can then be used to enrich perceptual object representations. With a minimal setup and an environment that enables puzzle-like tasks, we demonstrate the feasibility and benefits of this approach. Specifically, we compare three different methods of integrating semantic representations into a model-based RL architecture. Our experiments show that the degree of explicitness in knowledge separation correlates with faster learning, better accuracy, better generalization, and better interpretability.

Self-driving cars face complex driving situations with a large amount of agents when moving in crowded cities. However, some of the agents are actually not influencing the behavior of the self-driving car. Filtering out unimportant agents would inherently simplify the behavior or motion planning task for the system. The planning system can then focus on fewer agents to find optimal behavior solutions for the ego agent. This is helpful especially in terms of computational efficiency. In this paper, therefore, the research topic of importance filtering with driving risk models is introduced. We give an overview of state-of-the-art risk models and present newly adapted risk models for filtering. Their capability to filter out surrounding unimportant agents is compared in a large-scale experiment. As it turns out, the novel trajectory distance balances performance, robustness and efficiency well. Based on the results, we can further derive a novel filter architecture with multiple filter steps, for which risk models are recommended for each step, to further improve the robustness. We are confident that this will enable current behavior planning systems to better solve complex situations in everyday driving.

The survival analysis of driving trajectories allows for holistic evaluations of car-related risks caused by collisions or curvy roads. This analysis has advantages over common Time-To-X indicators, such as its predictive and probabilistic nature. However, so far, the theoretical risks have not been demonstrated in real-world environments. In this paper, we therefore present Risk Maps (RM) for online warning support in situations with forced lane changes, due to the end of roads. For this purpose, we first unify sensor data in a Relational Local Dynamic Map (R-LDM). RM is afterwards able to be run in real-time and efficiently probes a range of situations in order to determine risk-minimizing behaviors. Hereby, we focus on the improvement of uncertainty-awareness and transparency of the system. Risk, utility and comfort costs are included in a single formula and are intuitively visualized to the driver. In the conducted experiments, a low-cost sensor setup with a GNSS receiver for localization and multiple cameras for object detection are leveraged. The final system is successfully applied on two-lane roads and recommends lane change advices, which are separated in gap and no-gap indications. These results are promising and present an important step towards interpretable safety.

The task of driving autonomously is difficult due to the vast number of driving situations a system may be facing. Especially higher levels of automation in less restricted scopes remain a topic of active research. In previous work, we introduced a behavior planning system which uses analytic models to evaluate the quality of behavior holistically. It uses these models to generate quality-maximizing behavior instead of selecting among predefined behavior primitives. The system was able to solve various complex urban traffic scenarios in large-scale simulations. In this paper, we verify the system using multiple prototype vehicles on proving grounds in a number of difficult urban scenarios such as prioritized intersections or overtaking. We describe the system architecture and principles which render the system embodiment-agnostic and make extensions for additional features possible without massively increasing the complexity.

Many everyday traffic situations require cooperation among traffic participants to establish the order in which they pass a shared part of the road. Behavior planners which do not take this cooperative aspect into account properly struggle to find efficient solutions if the situation is nontrivial. Improper modelling may lead to overly aggressive or conservative behavior. In this paper, we propose an extension to state-of-the-art systems that enables behavior planners to efficiently cope with narrow passage scenarios even without car-to-car communication. The extended system is based on an asymmetry measure which takes the shared traffic space and the cooperation partners into account. This measure is then used to continuously predict which potential outcome is more likely to occur, to infer the assumed strategy of the cooperation partner, and to match the own strategy accordingly. Experiments show that the proposed system significantly reduces the cumulative passing time of the shared traffic space as compared to baseline systems. The resulting solutions are robust against variations in the behavior of both cooperation partners, and explicitly account for oblivious traffic participants which behave uncooperatively.

Estimation of Distribution Algorithms (EDAs) are metaheuristics where learning a model and sampling new solutions replaces the variation operators recombination and mutation used in standard Genetic Algorithms. The choice of these models as well as the corresponding training processes are subject to the bias/variance tradeoff, also known as under- and overfitting: simple models cannot capture complex interactions between problem variables, whereas complex models are susceptible to modeling random noise. This paper suggests using Denoising Autoencoders (DAEs) as generative models within EDAs (DAE-EDA). The resulting DAE-EDA is able to model complex probability distributions. Furthermore, overfitting is less harmful, since DAEs overfit by learning the identity function. This overfitting behavior introduces unbiased random noise into the samples, which is no major problem for the EDA but just leads to higher population diversity. As a result, DAE-EDA runs for more generations before convergence and searches promising parts of the solution space more thoroughly. We study the performance of DAE-EDA on several combinatorial single-objective optimization problems. In comparison to the Bayesian Optimization Algorithm, DAE-EDA requires a similar number of evaluations of the objective function but is much faster and can be parallelized efficiently, making it the preferred choice especially for large and difficult optimization problems.

We address the problem of motion planning for four-way intersection crossings with right-of-ways. Road safety typically assigns liability to the follower in rear-end collisions and to the approaching vehicle required to yield in side crashes. As an alternative to previous models based on heuristic state machines, we propose a planning framework which changes the prediction model of other cars (e.g. their prototypical accelerations and decelerations) depending on the given longitudinal or lateral priority rules. Combined with a state-of-the-art trajectory optimization approach ROPT (Risk Optimization Method) this allows to find ego velocity profiles minimizing risks from curves and all involved vehicles while maximizing utility (needed time to arrive at a goal) and comfort (change and duration of acceleration) under the presence of regulatory conditions. Analytical and statistical evaluations show that our method is able to follow right-of-ways for a wide range of other vehicle behaviors and path geometries. Even when the other cars drive in a non-priority-compliant way, ROPT achieves good risk-comfort tradeoffs.

Risk assessment is a central element for the development and validation of Autonomous Vehicles (AV). It comprises a combination of occurrence probability and severity of future critical events. Time Headway (TH) as well as Time-To-Contact (TTC) are commonly used risk metrics and have qualitative relations to occurrence probability. However, they lack theoretical derivations and additionally they are designed to only cover special types of traffic scenarios (e.g. following between single car pairs). In this paper, we present a probabilistic situation risk model based on survival analysis considerations and extend it to naturally incorporate sensory, temporal and behavioral uncertainties as they arise in real-world scenarios. The resulting Risk Spot Detector (RSD) is applied and tested on naturalistic driving data of a multi-lane boulevard with several intersections, enabling the visualization of road criticality maps. Compared to TH and TTC, our approach is more selective and specific in predicting risk. RSD concentrates on driving sections of high vehicle density where large accelerations and decelerations or approaches with high velocity occur.

We propose the set autoencoder, a model for unsupervised representation learning for sets of elements. It is closely related to sequence-to-sequence models, which learn fixed-sized latent representations for sequences, and have been applied to a number of challenging supervised sequence tasks such as machine translation, as well as unsupervised representation learning for sequences. In contrast to sequences, sets are permutation invariant. The proposed set autoencoder considers this fact, both with respect to the input as well as the output of the model. On the input side, we adapt a recently-introduced recurrent neural architecture using a content-based attention mechanism. On the output side, we use a stable marriage algorithm to align predictions to labels in the learning phase. We train the model on synthetic data sets of point clouds and show that the learned representations change smoothly with translations in the inputs, preserve distances in the inputs, and that the set size is represented directly. We apply the model to supervised tasks on the point clouds using the fixed-size latent representation. For a number of difficult classification problems, the results are better than those of a model that does not consider the permutation invariance. Especially for small training sets, the set-aware model benefits from unsupervised pretraining.

We consider the problem of correct motion planning for T-intersection merge-ins of arbitrary geometry and vehicle density. A merge-in support system has to estimate the chances that a gap between two consecutive vehicles can be taken successfully. In contrast to previous models based on heuristic gap size rules, we present an approach which optimizes the integral risk of the situation using parametrized velocity ramps. It accounts for the risks from curves and all involved vehicles (front and rear on all paths) with a so-called survival analysis. For comparison, we also introduce a specially designed extension of the Intelligent Driver Model (IDM) for entering intersections. We show in a quantitative statistical evaluation that the survival method provides advantages in terms of lower absolute risk (i.e., no crash happens) and better risk-utility tradeoff (i.e., making better use of appearing gaps). Furthermore, our approach generalizes to more complex situations with additional risk sources.

Estimation of Distribution Algorithms (EDAs) require flexible probability models that can be efficiently learned and sampled. Restricted Boltzmann Machines (RBMs) are generative neural networks with these desired properties. We integrate an RBM into an EDA and evaluate the performance of this system in solving combinatorial optimization problems with a single objective. We assess how the number of fitness evaluations and the CPU time scale with problem size and with problem complexity. The results are compared to the Bayesian Optimization Algorithm, a state-of-the-art EDA. Although RBM-EDA requires larger population sizes and a larger number of fitness evaluations, it outperforms BOA in terms of CPU times, in particular if the problem is large or complex. RBM-EDA requires less time for model building than BOA. These results highlight the potential of using generative neural networks for combinatorial optimization.

Estimation of Distribution Algorithms (EDAs) require flexible probability models that can be efficiently learned and sampled. Deep Boltzmann Machines (DBMs) are generative neural networks with these desired properties. We integrate a DBM into an EDA and evaluate the performance of this system in solving combinatorial optimization problems with a single objective. We compare the results to the Bayesian Optimization Algorithm. The performance of DBM-EDA was superior to BOA for difficult additively decomposable functions, i.e., concatenated deceptive traps of higher order. For most other benchmark problems, DBM-EDA cannot clearly outperform BOA, or other neural network-based EDAs. In particular, it often yields optimal solutions for a subset of the runs (with fewer evaluations than BOA), but is unable to provide reliable convergence to the global optimum competitively. At the same time, the model building process is computationally more expensive than that of other EDAs using probabilistic models from the neural network family, such as DAE-EDA.