In this talk, Elliot Weiss, Ph.D. in the Dynamic Design Lab will present his work on how to conduct immersive experiments of novel shared control architectures and how to apply shared control to scenarios involving sequences of distinct maneuvers.
Devansh Jalota, PhD candidate in Marco Pavone's Autonomous Systems Lab presents presents his work on congestion pricing mechanisms to achieve the efficiency and equity objectives of sustainable transportation.
Sonia Martin, PhD candidate in Ram Rajagopal's Stanford Sustainable Systems Lab presents her research on green charging, or smart charging control, a control strategy based on actual EV driver data and historical grid emissions.
Anthony Corso, postdoctoral researcher in the Stanford Intelligent Systems Lab presented two promising approaches for safety validation: formal verification of neural networks and black-box adaptive sampling in the context of aviation and driving applications. Together these techniques may provide a path forward to ensuring the safe deployment of autonomous systems.
Human play data consists of unstructured, reset-free interactions with the environment, where the human is able to choose both the tasks and how the tasks are executed. Suneel Belkhale with the Stanford Intelligent and Interactive Autonomous Systems Group (ILIAD) discusses methods for learning from human play data and presents the lab's recent method - PLATO: Predicting Latent Affordances Through Object-centric Play - which views play data as sequential object interactions in order to learn a robust policy from play.
Knowledge of the preferences and actions of other users is essential for making optimal traffic routing decisions. However, in practice, this information may not be available due to privacy considerations, which makes it challenging to compute efficient solutions. In this talk, CARS Fellow, Karthik Gopalakrishnan, will describe two settings where we achieve efficient routing without compromising user privacy by leveraging ideas from cryptography, differential privacy, and online learning.
Justin Luke, PhD candidate in Civil and Environmental Engineering presents an optimization model that jointly determines the vehicle sizing, the siting of charging stations of varying charging speeds, and the macroscopic-level operations for an autonomous electric mobility-on-demand fleet. Results from a case-study in San Francisco showing the type of EVs and optimal siting of charging stations and sizes are also presented.
Autonomous vehicles should be able to maintain control in scenarios that push them beyond the traditional limits of handling. Especially in situations when the rear tires are sliding, and the vehicle is drifting because the natural dynamics become unstable and can cause the vehicle to spin, endangering the safety of the occupants and others on the road. In this webinar we’ll discuss how we can study professional drivers and improve the state of the art of autonomous drifting controllers.
To navigate complex driving scenarios, automated vehicles must be able to make decisions that reflect higher-level goals such as safety and efficiency, leveraging the vehicle’s full capabilities if necessary. We introduce an architecture that is capable of handling combinatorial decision making and control with a high fidelity vehicle model. This is accomplished by solving a nonlinear model predictive control optimization for each maneuver variant, or homotopy, identified in the drivable space.
To reap the rewards smart mobility holds with efficient, sustainable and equitable urban transportation, we must develop control mechanisms capable of tackling massive scenarios while accoutning for their societal impact in terms of congestion, pollution and fairness, to mention just a few criteria. In this talk, Kiril Solovey, post doctoral scholar with the Autonomous Systems Lab, describes some fundamental ideas on the control of large-sale transportation systems and broad implications of such systems to society and highlight directions for future research.
