Effects of Partial Observability Solver Methods on Training and Final Policies in Autonomous Driver RL: How do different methods for dealing with partial observability in the environment influence training and the robustness of final policies under various testing conditions?

Autonomous driving is a complex problem that can potentially be solved using artificial intelligence. The complexity stems from the system's need to understand the surroundings and make appropriate decisions. However, there are various challenges in constructing such a sophisticated system. One of the main challenges is to make the agent learn...

Evaluating Robustness of Deep Reinforcement Learning for Autonomous Driving: How does entropy maximization affect the training and robustness of final policies under various testing conditions?

This research paper aims to investigate the effect of entropy while training the agent on the robustness of the agent. This is important because robustness is defined as the agent's adaptability to different environments. A self-driving car should adapt to every environment that it is being used in since a mistake could cost someone's life....

A user centered vision for the ecosystem of shared autonomous vehicles: For 2040 within the MRDH

Our personal mobility systems will undergo a significant transformation with the emergence of electrified drivetrains, shared services, and autonomous technology (Sprei, 2018). We can already plan door-to-door transport, share various mobility modes, drive fully electric cars, and several vehicles can assist with some driving functions like...

How Human-Machine Interaction keeps pace with Automated Vehicles: A systematic review

Human Machine Interface (HMI) is a design concept that improves the interaction between the driver and the automated vehicle, which leads to greater safety and comfort for the driver and greater safety for the road user. Therefore, many papers and patents are published every year. Many papers use different methodologies and materials due to some...

Plastic intensive, purpose-built car: Autonomous urban transport in 2029

It is expected that the automotive industry will be facing a radical shift, going from human driven cars to cars that are driving autonomously. In this project I started out with performing an in depth literature research to get a good understanding on what is expected to change. My expectations that the design and solutions for a self-driving...

Paving the way for autonomous cars: Current projects and challenges in the Netherlands

<br/>Currently, cities are in the midst of a transportation transition, namely, from human-driven to driverless vehicles. Because of the rapid development of technology, many systems are being automated or even becoming autonomous. Autonomous means that the system is able to perform each task automatically without any human intervention. Much...

An autonomous urban delivery concept for Ford

Ordering things online is increasing in popularity. This causes logistical problems in the urban space, such as congestion, emissions and other irritations. As autonomous vehicle technology is becoming mature, a new opportunity arises: autonomous parcel delivery. During this graduation assignment with Ford, an autonomous urban delivery system...

Learning to Control Multi- Dimensional Autonomous Agents using Hebbian Learning: A Global Reward Approach

The novelty-raahn algorithm has been shown to effectively learn a desired behavior from raw inputs by connecting an autoencoder with a Hebbian network. Hebbian learning is compelling for its biological plausibility and simplicity. It changes the weight of a connection based only on the activations of neurons it connects, and can effectively...

Reactive Environment Outlet: Shape morphing application in autonomous cars

This was an explorative project which investigated the use and implementation of smart materials in an automotive context for Toyota Motors Europe. The smart materials in this context were shape memory materials which possess the ability to change their shape under appropriate stimulus. The idea of the project was to understand how the materials...

Testing Autonomous Cars for Feature Interaction Failures using Many-Objective Search

Complex systems such as autonomous cars are typically built as a composition of features that are independent units of functionality. Features tend to interact and impact one another’s behavior in unknown ways. A challenge is to detect and manage feature interactions, in particular, those that violate system requirements, hence leading to...

Compositional and contract-based verification for autonomous driving on road networks

Recent advances in autonomous driving have raised the problem of safety<br/>to the forefront and incentivized research into establishing safety guarantees. In this paper, we propose a safety verification framework as a safety standard for driving controllers with full or shared autonomy based on compositional and contract-based principles. Our...

LTL specifications for Highway Lane Changing Maneuvers of Highly Automated Vehicles

Since the interest in autonomous driving solutions is massively increasing, the need for good and reliable control algorithms is growing every day. This project studies the performance of safe lane changes of a highly autonomous vehicle given the currently available perception of the environment, vehicle dynamics and desired comfort and speed...