People as Sensors: Imputing Maps from Human Actions

Afolabi, Oladapo; Driggs-Campbell, Katherine; Dong, Roy; Kochenderfer, Mykel J.; Sastry, S. Shankar

doi:10.1109/iros.2018.8594511

Cited by 28 publications

(27 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These problems are orthogonal to the problem of occluded factor inference from the behaviour of an observed vehicle. The works most closely related to ours are the works of Sun et al [29] and Afolabi et al [1] who infer occluded obstacles from the actions of observed vehicles. Sun et al [29] infer so-called social information (i.e., occluded factors) with an inverse-planning approach based on a reward function learned with inverse reinforcement learning.…”

Section: Related Workmentioning

confidence: 79%

Interpretable Goal Recognition in the Presence of Occluded Factors for Autonomous Vehicles

Hanna¹,

Rahman²,

Fosong³

et al. 2021

Preprint

View full text Add to dashboard Cite

Recognising the goals or intentions of observed vehicles is a key step towards predicting the long-term future behaviour of other agents in an autonomous driving scenario. When there are unseen obstacles or occluded vehicles in a scenario, goal recognition may be confounded by the effects of these unseen entities on the behaviour of observed vehicles. Existing prediction algorithms that assume rational behaviour with respect to inferred goals may fail to make accurate long-horizon predictions because they ignore the possibility that the behaviour is influenced by such unseen entities. We introduce the Goal and Occluded Factor Inference (GOFI) algorithm which bases inference on inverse-planning to jointly infer a probabilistic belief over goals and potential occluded factors. We then show how these beliefs can be integrated into Monte Carlo Tree Search (MCTS). We demonstrate that jointly inferring goals and occluded factors leads to more accurate beliefs with respect to the true world state and allows an agent to safely navigate several scenarios where other baselines take unsafe actions leading to collisions.

show abstract

Section: Related Workmentioning

confidence: 79%

Interpretable Goal Recognition in the Presence of Occluded Factors for Autonomous Vehicles

Hanna¹,

Rahman²,

Fosong³

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Similarly, several studies have attempted to predict the states of objects in blind spots in traffic scenes. Afolabi et al [15] estimated occupancy grids of pedestrians in blind spots based on visible driver behaviors. Sun et al [16] proposed a planning framework for driving behaviors using human traffic participants as distributed sensors.…”

Section: B Prediction Of Invisible Objectsmentioning

confidence: 99%

Predicting Appearance of Vehicles From Blind Spots Based on Pedestrian Behaviors at Crossroads

Hara

Kataoka

Inaba

et al. 2022

IEEE Trans. Intell. Transport. Syst.

View full text Add to dashboard Cite

Conventional prediction approaches for traffic scenes primarily predict the future states of visible objects (i.e., not in blind spots) based on their current observations. This study focused on the prediction of future states of objects in blind spots (e.g., those outside the filed-of-view or occluded regions) based on the current observations of other visible objects. We proposed a method that predicts the appearance of vehicles from a blind spot based on the behaviors of visible pedestrians who observe vehicles in the blind spot. Our proposed method utilizes a spatiotemporal 3D convolutional neural network and learns pedestrian behaviors for predictions. The method explicitly represents subtle motions and the surrounding environments of pedestrians using pose estimation and semantic segmentation. To conduct evaluation experiments, we built two datasets of videos capturing real traffic scenes. The datasets are collected by cameras with and without ego-motions. Using the datasets, we conducted experiments not only on simpler configurations but also on realistic traffic environments. Based on the experimental results, the following conclusions could be obtained: (i) our proposed method achieved a high performance at a level similar to that of humans in our prediction task, and predicted the appearance of vehicles from blind spots more than 1.5 s before they actually appeared. (ii) Explicit representations of pose and semantic masks captured information complementary to RGB videos, and ensembling the representations improved the prediction performance. (iii) Fine-tuning the models using videos with ego-motions is important to achieve good prediction in the videos captured by driving cars.

show abstract

“…Because all vehicles cannot be equipped with vehicle to vehicle (V2V) communications at present, autonomous vehicles should perceive the surrounding environment based on local sensors. Social perception has been devised to deal with local sensor limits [33], [34]. The perceived vehicle information could be used to infer the area beyond the blind spot or sensor limit.…”

Section: Introductionmentioning

confidence: 99%

Virtual Target-Based Overtaking Decision, Motion Planning, and Control of Autonomous Vehicles

Chae

2020

IEEE Access

View full text Add to dashboard Cite

This paper describes the design, implementation, and evaluation of a virtual target-based overtaking decision, motion planning, and control algorithm for autonomous vehicles. Both driver acceptance and safety, when surrounded by other vehicles, must be considered during autonomous overtaking. These are considered through safe distance based on human driving behavior. Since all vehicles cannot be equipped with a vehicle to vehicle communications at present, autonomous vehicles should perceive the surrounding environment based on local sensors. In this paper, virtual targets are devised to cope with the limitation of cognitive range. A probabilistic prediction is adopted to enhance safety, given the potential behavior of surrounding vehicles. Then, decision-making and motion planning has been designed based on the probabilistic prediction-based safe distance, which could achieve safety performance without a heavy computational burden. The algorithm has considered the decision rules that drivers use when overtaking. For this purpose, concepts of target space, demand, and possibility for lane change are devised. In this paper, three driving modes are developed for active overtaking. The desired driving mode is decided for safe and efficient overtaking. To obtain desired states and constraints, intuitive motion planning is conducted. A stochastic model predictive control has been adopted to determine vehicle control inputs. The proposed autonomous overtaking algorithm has been evaluated through simulation, which reveals the effectiveness of virtual targets. Also, the proposed algorithm has been successfully implemented on an autonomous vehicle and evaluated via real-world driving tests. Safe and comfortable overtaking driving has been demonstrated using a test vehicle. INDEX TERMS Autonomous vehicles, autonomous driving, decision-making, motion planning, vehicle control, overtaking, lane change, virtual target.

show abstract

People as Sensors: Imputing Maps from Human Actions

Cited by 28 publications

References 20 publications

Interpretable Goal Recognition in the Presence of Occluded Factors for Autonomous Vehicles

Interpretable Goal Recognition in the Presence of Occluded Factors for Autonomous Vehicles

Predicting Appearance of Vehicles From Blind Spots Based on Pedestrian Behaviors at Crossroads

Virtual Target-Based Overtaking Decision, Motion Planning, and Control of Autonomous Vehicles

Contact Info

Product

Resources

About