Forecasting the motion of surrounding vehicles is a critical ability for an autonomous vehicle deployed in complex traffic. Motion of all vehicles in a scene is governed by the traffic context, i.e., the motion and relative spatial configuration of neighboring vehicles. In this paper we propose an LSTM encoder-decoder model that uses convolutional social pooling as an improvement to social pooling layers for robustly learning interdependencies in vehicle motion. Additionally, our model outputs a multi-modal predictive distribution over future trajectories based on maneuver classes. We evaluate our model using the publicly available NGSIM US-101 and I-80 datasets. Our results show improvement over the state of the art in terms of RMS values of prediction error and negative log-likelihoods of true future trajectories under the model's predictive distribution. We also present a qualitative analysis of the model's predicted distributions for various traffic scenarios.

/pdf/convolutional-social-pooling-for-vehicle-trajectory-4fzlp3b04v.pdf

Convolutional Social Pooling for Vehicle Trajectory Prediction

To safely and efficiently navigate through complex traffic scenarios, autonomous vehicles need to have the ability to predict the future motion of surrounding vehicles. Multiple interacting agents, the multi-modal nature of driver behavior, and the inherent uncertainty involved in the task make motion prediction of surrounding vehicles a challenging problem. In this paper, we present an LSTM model for interaction aware motion prediction of surrounding vehicles on freeways. Our model assigns confidence values to maneuvers being performed by vehicles and outputs a multi-modal distribution over future motion based on them. We compare our approach with the prior art for vehicle motion prediction on the publicly available NGSIM US-101 and I-80 datasets. Our results show an improvement in terms of RMS values of prediction error. We also present an ablative analysis of the components of our proposed model and analyze the predictions made by the model in complex traffic scenarios.

Multi-Modal Trajectory Prediction of Surrounding Vehicles with Maneuver based LSTMs

A vehicle control device includes a detecting unit that detects a towing state of a subject vehicle, a recognizing unit that recognizes surrounding situations of the subject vehicle, a control unit that performs automatic control in which at least one of acceleration-and-deceleration and steering of the subject vehicle is automatically controlled on the basis of the surrounding situations of the subject vehicle recognized by the recognizing unit, and a changing unit that changes, if the detecting unit has detected that the subject vehicle is in a state of towing an object, details of control performed by the control unit in such a manner that it is less likely to perform the automatic control than in the case where the detecting unit has not detected that the subject vehicle is in a state of towing an object

Vehicle control system, vehicle control method, and vehicle control program

Self-driving vehicles will soon be a reality, as main automotive companies have announced that they will sell their driving automation modes in the 2020s. This technology raises relevant controversies, especially with recent deadly accidents. Nevertheless, autonomous vehicles are still popular and attractive thanks to the improvement they represent to people’s way of life (safer and quicker transit, more accessible, comfortable, convenient, efficient, and environment-friendly). This paper presents a review of motion planning techniques over the last decade with a focus on highway planning. In the context of this article, motion planning denotes path generation and decision making. Highway situations limit the problem to high speed and small curvature roads, with specific driver rules, under a constrained environment framework. Lane change, obstacle avoidance, car following, and merging are the situations addressed in this paper. After a brief introduction to the context of autonomous ground vehicles, the detailed conditions for motion planning are described. The main algorithms in motion planning, their features, and their applications to highway driving are reviewed, along with current and future challenges and open issues.

/pdf/a-review-of-motion-planning-for-highway-autonomous-driving-56ou2tzah8.pdf

A Review of Motion Planning for Highway Autonomous Driving

Reliable prediction of surround vehicle motion is a critical requirement for path planning for autonomous vehicles. In this paper we propose a unified framework for surround vehicle maneuver classification and motion prediction that exploits multiple cues, namely, the estimated motion of vehicles, an understanding of typical motion patterns of freeway traffic and inter-vehicle interaction. We report our results in terms of maneuver classification accuracy and mean and median absolute error of predicted trajectories against the ground truth for real traffic data collected using vehicle mounted sensors on freeways. An ablative analysis is performed to analyze the relative importance of each cue for trajectory prediction. Additionally, an analysis of execution time for the components of the framework is presented. Finally, we present multiple case studies analyzing the outputs of our model for complex traffic scenarios

How would surround vehicles move? A Unified Framework for Maneuver Classification and Motion Prediction

Advanced driver assistance systems or highly automated driving systems for lane change maneuvers are expected to enhance highway traffic safety, transport efficiency, and driver comfort. To extend the capability of current advanced driver assistance systems, and eventually progress to highly automated highway driving, the task of automatically determine if, when, and how to perform a lane change maneuver, is essential. This paper thereby presents a low-complexity lane change maneuver algorithm which determines whether a lane change maneuver is desirable, and if so, selects an appropriate inter-vehicle traffic gap and time instance to perform the maneuver, and calculates the corresponding longitudinal and lateral control trajectory. The ability of the proposed lane change maneuver algorithm to make appropriate maneuver decisions and generate smooth and safe lane change trajectories in various traffic situations is demonstrated by simulation and experimental results.

If, When, and How to Perform Lane Change Maneuvers on Highways

A driver intention estimation arrangement is provided to estimate a level of overtaking intention of a driver of a vehicle hosting the arrangement. The arrangement includes a driver eye monitoring unit configured to monitor a gaze direction (g) of at least one eye of the driver, and a processing unit configured to analyze the gaze direction (g). The processing unit is configured to estimate the level of overtaking intention on the basis of the gaze direction (g). A forward collision warning system, an autonomous brake system, a vehicle, and a method of estimating a level of overtaking intention of a host vehicle driver are also provided.

Driver intention estimation arrangement

The present disclosure relates to a method performed by an evasive manoeuvre system (1) for providing a driver behaviour adapted evasive manoeuvre to a vehicle (2) at risk of an impending or probable collision. The evasive manoeuvre system detects (1001) a driving environment (3). Furthermore, the evasive manoeuvre system determines (1002) that the vehicle is at risk of colliding with an obstacle (5). The evasive manoeuvre system moreover determines (1003) a drivable zone (6) considered safe driving for the vehicle. Further, the evasive manoeuvre system detects (1004) a driver initiated collision avoidance manoeuvre. Moreover, the evasive manoeuvre system intervenes (1005) in control of the vehicle, such that the vehicle is maintained within the drivable zone. The disclosure also relates to an evasive manoeuvre system in accordance with the foregoing, and a vehicle comprising at least a portion of the evasive manoeuvre system.

Method and system for providing a driver behaviour adapted evasive manoeuvre

A method is disclosed for suppressing or deactivating a lane keeping assist (LKA) system in a host vehicle. The method may include retrieving environmental data from a data gathering system, determining a current lane width based on the environmental data retrieved from the data gathering system, and comparing the current lane width with a pre-set lane width. The method may also include determining that the current lane width is smaller than the pre-set lane width, detecting the presence of an obstacle on the traveling path which obstacle will be passed by the host vehicle or which obstacle will pass the host vehicle within a pre-determined time period, and deactivating or suppressing the lane keeping assist (LKA) system.

Lane keeping suppressing system and method

The present invention relates to a method for providing an object prediction representation, which method i.a. comprises simulating movement of a host vehicle and an object in a longitudinal-lateral coordinate system based on their initial positions at the starting time (t0) and a movement of the host vehicle and determined movement of the object, wherein points in time (t0-n) when at least one reference point of the host vehicle and an object reference point of a plurality of object reference points have the same longitudinal position are detected, and wherein for each detected point in time an associated lateral position (y1-n) of the object reference point at the detected point in time is detected.

Jonatan Silvlin

Papers

If, When, and How to Perform Lane Change Maneuvers on Highways

Driver intention estimation arrangement

Method and system for providing a driver behaviour adapted evasive manoeuvre

Lane keeping suppressing system and method

Method for providing an object prediction representation