Search

There is a growing need for open source hardware subcomponents to be evaluated. Most robotic systems are ultimately based upon motors which are driven to move either to certain positions, as in robot arms, or to certain velocities, as in wheeled mobile robots. We evaluate a state of the art OSH driver, OSMC, for such systems, and contribute new Open Source Software (OSS) to control it. Our findings suggest that OSMC is now mature enough to replace closed-source motor drivers in medium-size robots such as agri-robots and last mile delivery vehicles.

Human-robot interaction (HRI) requires quantitative models of proxemics and trust for robots to use in negotiating with people for space. Hall’s theory of proxemics has been used for decades to describe social interaction distances but has lacked detailed quantitative models and generative explanations to apply to these cases. In the limited case of autonomous vehicle interactions with pedestrians crossing a road, a recent model has explained the quantitative sizes of Hall’s distances to 4% error and their links to the concept of trust in human interactions. The present study extends this model by generalising several of its assumptions to cover further cases including human-human and human-robot interactions. It tightens the explanations of Hall zones from 4% to 1% error and fits several more recent empirical HRI results. This may help to further unify these disparate fields and quantify them to a level which enables real-world operational HRI applications.

To be successful, automated vehicles (AVs) need to be able to manoeuvre in mixed traffic in a way that will be accepted by road users, and maximises traffic safety and efficiency. A likely prerequisite for this success is for AVs to be able to commu- nicate effectively with other road users in a complex traffic environment. The current study, conducted as part of the European project interACT, investigates the communication strategies used by drivers and pedestrians while crossing the road at six observed locations, across three European countries. In total, 701 road user interactions were observed and annotated, using an observation protocol developed for this purpose. The observation protocols identified 20 event categories, observed from the approaching vehicles/drivers and pedestrians. These included information about movement, looking behaviour, hand gestures, and signals used, as well as some demographic data. These observations illustrated that explicit communication techniques, such as honking, flashing headlights by drivers, or hand gestures by drivers and pedestrians, rarely occurred. This observation was consistent across sites. In addition, a follow-on questionnaire, administered to a sub-set of the observed pedestrians after crossing the road, found that when contemplating a crossing, pedestrians were more likely to use vehicle- based behaviour, rather than communication cues from the driver. Overall, the findings suggest that vehicle-based movement information such as yielding cues are more likely to be used by pedestrians while crossing the road, compared to explicit communication cues from drivers, although some cultural differences were observed. The implications of these findings are discussed with respect to design of suitable external interfaces and communication of intent by future automated vehicles.

Autonomous vehicles now have well developed algorithms and open source software for localisation and navigation in static environments but their future interactions with other road users in mixed traffic environments, especially with pedestrians, raise some concerns. Pedestrian behaviour is complex to model and unpredictable, thus creating a big challenge for self-driving cars. This paper examines pedestrian behaviour during crossing scenarios with a game theoretic autonomous vehicle in virtual reality. In a first experiment, we recorded participants’ trajectories and found that they were crossing more cautiously in VR than in previous laboratory experiments. In two other experiments, we used a gradient descent approach to investigate participants’ preference for a certain AV driving style. We found that the majority of them were not expecting the car to stop in these scenarios. These results suggest that VR is an interesting tool for testing autonomous vehicle algorithms and for finding out about pedestrian preferences.

Robot navigation in environments with static objects appears to be a solved problem, but navigating around humans in dynamic and unstructured environments remains an active research question. This requires not only advanced path planning methods but also a good perception system, models of multi-agent interactions and realistic hardware for testing. To evolve in human social spaces, robots must also show social intelligence, i.e. the ability to understand human behaviour via explicit and implicit communication cues (e.g. proxemics) for better human-robot interactions (HRI) [28]. Similarly, autonomous vehicles (AVs), also called “self-driving cars” that are appearing on the roads need a better understanding of pedestrians’ social behaviour, especially in urban areas [26]. In particular, previous work showed that pedestrians may take advantage over autonomous vehicles [13] by intentionally and constantly stepping in front of AVs, hence preventing them from making progress on the roads. This inability of current AVs to read the intention of other road users, predict their future behaviour and interact with them is known as “the big problem with self-driving cars” [1]. Thus, AVs need better decision-making models and must find a good balance between stopping for pedestrians when required and driving to reach their final destination as quickly as possible for their on-board passengers. A comprehensive review of existing pedestrian models for AVs, ranging from low-level sensing, detection and tracking models [9] to high-level interaction and game theoretic models of pedestrian behaviour [10], found that the lower-level models are accurate and mature enough to be deployed on AVs but more research is needed in the higher-level models. Hence, in this work, we focus on modelling, learning and operating pedestrian high-level social behaviour on self-driving cars using game theory and proxemics.

Recent years have witnessed the rapid deployment of robotic systems in public places such as roads, pavements, workplaces and care homes. Robot navigation in environments with static objects is largely solved, but navigating around humans in dynamic environments remains an active research question for autonomous vehicles (AVs). To navigate in human social spaces, self-driving cars and other robots must also show social intelligence. This involves predicting and planning around pedestrians, understanding their personal space, and establishing trust with them. Most current AVs, for legal and safety reasons, consider pedestrians to be obstacles, so these AVs always stop for or replan to drive around them. But this highly safe nature may lead pedestrians to take advantage over them and slow their progress, even to a complete halt. We provide a review of our recent research on predicting and controlling human–AV interactions, which combines game theory, proxemics and trust, and unifies these fields via quantitative, probabilistic models and robot controllers, to solve this “freezing robot” problem.

Autonomous Vehicles (AVs) must interact with other road users. They must understand and adapt to complex pedestrian behaviour, especially during crossings where priority is not clearly defined. This includes feedback effects such as modelling a pedestrian’s likely behaviours resulting from changes in the AVs behaviour. For example, whether a pedestrian will yield if the AV accelerates, and vice versa. To enable such automated interactions, it is necessary for the AV to possess a statistical model of the pedestrian’s responses to its own actions. A previous work demonstrated a proof-of- concept method to fit parameters to a simplified model based on data from a highly artificial discrete laboratory task with human subjects. The method was based on LIDAR-based person tracking, game theory, and Gaussian process analysis. The present study extends this method to enable analysis of more realistic continuous human experimental data. It shows for the first time how game-theoretic predictive parameters can be fit into pedestrians natural and continuous motion during road-crossings, and how predictions can be made about their interactions with AV controllers in similar real-world settings.

Localisation and navigation of autonomous vehicles (AVs) in static environments are now solved problems, but how to control their interactions with other road users in mixed traffic environments, especially with pedestrians, remains an open question. Recent work has begun to apply game theory to model and control AV-pedestrian interactions as they compete for space on the road whilst trying to avoid collisions. But this game theory model has been developed only in unrealistic lab environments. To improve their realism, this study empirically examines pedestrian behaviour during road crossing in the presence of approaching autonomous vehicles in more realistic virtual reality (VR) environments. The autonomous vehicles are controlled using game theory, and this study seeks to find the best parameters for these controls to produce comfortable interactions for the pedestrians. In a first experiment, participants’ trajectories reveal a more cautious crossing behaviour in VR than in previous laboratory experiments. In two further experiments, a gradient descent approach is used to investigate participants’ preference for AV driving style. The results show that the majority of participants were not expecting the AV to stop in some scenarios, and there was no change in their crossing behaviour in two environments and with different car models suggestive of car and last-mile style vehicles. These results provide some initial estimates for game theoretic parameters needed by future AVs in their pedestrian interactions and more generally show how such parameters can be inferred from virtual reality experiments.

—Autonomous vehicles (AVs) must share space with pedestrians, both in carriageway cases such as cars at pedestrian crossings and off-carriageway cases such as delivery vehicles navigating through crowds on pedestrianized high-streets. Unlike static obstacles, pedestrians are active agents with complex, inter- active motions. Planning AV actions in the presence of pedestrians thus requires modelling of their probable future behaviour as well as detecting and tracking them. This narrative review article is Part I of a pair, together surveying the current technology stack involved in this process, organising recent research into a hierarchical taxonomy ranging from low-level image detection to high-level psychology models, from the perspective of an AV designer. This self-contained Part I covers the lower levels of this stack, from sensing, through detection and recognition, up to tracking of pedestrians. Technologies at these levels are found to be mature and available as foundations for use in high-level systems, such as behaviour modelling, prediction and interaction control.

Understanding pedestrian behaviour and controlling interactions with pedestrians is of critical importance for autonomous vehicles, but remains a complex and challenging problem. This study infers pedestrian intent during possible road-crossing interactions, to assist autonomous vehicle decisions to yield or not yield when approaching them, and tests a simple heuristic model of intent on pedestrian-vehicle trajectory data for the first time. It relies on a heuristic approach based on the observed positions of the agents over time. The method can predict pedestrian crossing intent, crossing or stopping, with 96% accuracy by the time the pedestrian reaches the curbside, on the standard Daimler pedestrian dataset. This result is important in demarcating scenarios which have a clear winner and can be predicted easily with the simple heuristic, from those which may require more complex game-theoretic models to predict and control.

Interacting with humans remains a challenge for autonomous vehicles (AVs). When a pedestrian wishes to cross the road in front of the vehicle at an unmarked crossing, the pedestrian and AV must compete for the space, which may be considered as a game-theoretic interaction in which one agent must yield to the other. To inform development of new real-time AV controllers in this setting, this study collects and analy- ses detailed, manually-annotated, temporal data from real-world human road crossings as they interact with manual drive vehicles. It studies the temporal orderings (filtrations) in which features are revealed to the ve- hicle and their informativeness over time. It presents a new framework suggesting how optimal stopping controllers may then use such data to enable an AV to decide when to act (by speeding up, slowing down, or otherwise signalling intent to the pedestrian) or alternatively, to continue at its current speed in order to gather additional information from new features, including signals from that pedestrian, before acting itself.