Adaptive Trajectory Generation of Autonomous Vehicle in Urban Environments

This work is intended to give an overview of technologies, developed from an artificial intelligence standpoint, devised to face the different planning and control problems involved in trajectory generation for mobile robots. The purpose of this analysis is to give a current context to present the Evolutionary Robotics approach to the problem, which is now being considered as a feasible methodology to develop mobile robots for solving real life problems. This chapter also show the authors’ experiences on related case studies, which are briefly described (a fuzzy logic based path planner for a terrestrial mobile robot, and a knowledge-based system for desired trajectory generation in the Geosub underwater autonomous vehicle). The development of different behaviours within a path generator, built with Evolutionary Robotics concepts, is tested in a Khepera© robot and analyzed in detail. Finally, behaviour coordination based on the artificial immune system metaphor is evaluated for the same application.

Download Full-text

1101 Development of an Algorithm for the Trajectory Generation of a Control Target in a Driver Model and Vehicle Dynamics for an Autonomous Vehicle

The Proceedings of the Transportation and Logistics Conference ◽

10.1299/jsmetld.2011.20.131 ◽

2011 ◽

Vol 2011.20 (0) ◽

pp. 131-134

Author(s):

Satoshi Sugiyama ◽

Tetsuya Kaneko ◽

Yukiyo Kuriyagawa ◽

Ichiro Kageyama

Keyword(s):

Vehicle Dynamics ◽

Autonomous Vehicle ◽

Trajectory Generation ◽

Driver Model ◽

Control Target

Download Full-text

Deep tracking in the wild: End-to-end tracking using recurrent neural networks

The International Journal of Robotics Research ◽

10.1177/0278364917710543 ◽

2017 ◽

Vol 37 (4-5) ◽

pp. 492-512 ◽

Cited By ~ 23

Author(s):

Julie Dequaire ◽

Peter Ondrúška ◽

Dushyant Rao ◽

Dominic Wang ◽

Ingmar Posner

Keyword(s):

Object Tracking ◽

Autonomous Vehicle ◽

Urban Environments ◽

Positive Contribution ◽

Efficient Manner ◽

Model Free ◽

Novel Approach ◽

In The Wild ◽

End To End ◽

Traditional Approaches

This paper presents a novel approach for tracking static and dynamic objects for an autonomous vehicle operating in complex urban environments. Whereas traditional approaches for tracking often feature numerous hand-engineered stages, this method is learned end-to-end and can directly predict a fully unoccluded occupancy grid from raw laser input. We employ a recurrent neural network to capture the state and evolution of the environment, and train the model in an entirely unsupervised manner. In doing so, our use case compares to model-free, multi-object tracking although we do not explicitly perform the underlying data-association process. Further, we demonstrate that the underlying representation learned for the tracking task can be leveraged via inductive transfer to train an object detector in a data efficient manner. We motivate a number of architectural features and show the positive contribution of dilated convolutions, dynamic and static memory units to the task of tracking and classifying complex dynamic scenes through full occlusion. Our experimental results illustrate the ability of the model to track cars, buses, pedestrians, and cyclists from both moving and stationary platforms. Further, we compare and contrast the approach with a more traditional model-free multi-object tracking pipeline, demonstrating that it can more accurately predict future states of objects from current inputs.

Download Full-text

Accurate feature matching for autonomous vehicle navigation in urban environments

2011 16th International Conference on Methods & Models in Automation & Robotics ◽

10.1109/mmar.2011.6031318 ◽

2011 ◽

Cited By ~ 3

Author(s):

Jurek Z. Sasiadek ◽

Mark J. Walker ◽

Adam Krzyzak

Keyword(s):

Feature Matching ◽

Autonomous Vehicle ◽

Urban Environments ◽

Vehicle Navigation ◽

Autonomous Vehicle Navigation

Download Full-text

The Value of Autonomous Vehicles for Last-Mile Deliveries in Urban Environments

Management Science ◽

10.1287/mnsc.2020.3917 ◽

2021 ◽

Author(s):

Sara Reed ◽

Ann Melissa Campbell ◽

Barrett W. Thomas

Keyword(s):

Operations Management ◽

Autonomous Vehicles ◽

Completion Time ◽

Autonomous Vehicle ◽

Optimal Solution ◽

Fixed Time ◽

Urban Environments ◽

Assisted Delivery ◽

Autonomous Case ◽

Number Of Customers

We demonstrate that autonomous-assisted delivery can yield significant improvements relative to today’s system in which a delivery person must park the vehicle before delivering packages. We model an autonomous vehicle that can drop off the delivery person at selected points in the city where the delivery person makes deliveries to the final addresses on foot. Then, the vehicle picks up the delivery person and travels to the next reloading point. In this way, the delivery person would never need to look for parking or walk back to a parking place. Based on the number of customers, driving speed of the vehicle, walking speed of the delivery person, and the time for loading packages, we characterize the optimal solution to the autonomous case on a solid rectangular grid of customers, showing the optimal solution can be found in polynomial time. To benchmark the completion time of the autonomous case, we introduce a traditional model for package delivery services that includes the time to search for parking. If the time to find parking is ignored, we show the introduction of an autonomous vehicle reduces the completion time of delivery to all customers by 0%–33%. When nonzero times to find parking are considered, the delivery person saves 30%–77% with higher values achieved for longer parking times, smaller capacities, and lower fixed time for loading packages. This paper was accepted by Vishal Gaur, operations management.

Download Full-text