DAY 2: Main Symposium

We describe recent work on coordinating large quadrotor teams in obstacle-rich environments. We begin with a method for formation-change trajectory planning. Our method decomposes the planning problem into two stages: a discrete planner operating on a graph representation of the workspace, and a continuous refinement that converts the non-smooth graph plan into a set of Ck-continuous trajectories, locally optimizing an integral-squared-derivative cost. We account for the downwash effect, allowing safe flight in dense formations. We demonstrate computational efficiency in simulation with up to 200 robots and physical plausibility with an experiment with 32 small quadrotors. Our approach computes safe and smooth trajectories for hundreds of quadrotors in dense environments with obstacles in a few minutes. We have recently begun a project to investigate whether it is possible to learn controllers for such behavior via large-scale multi-agent end-to-end reinforcement learning. Our recent results suggest that it is possible to train policies parameterized by neural networks that are capable of controlling individual drones in a swarm in a fully decentralized manner. Our policies, trained in simulated environments with realistic quadrotor physics, demonstrate advanced flocking behaviors, perform aggressive maneuvers in tight formations while avoiding collisions with each other, break and re-establish formations to avoid collisions with moving obstacles, and efficiently coordinate in pursuit-evasion tasks.

Trading computation and decision making for less communication, sensing, or actuator effort offers great promise for the autonomous operation of both individual and interconnected cyberphysical systems. Resource-aware control seeks to prescribe, in a principled way, when to use the available resources efficiently while still guaranteeing a desired quality of service in performing the intended task. This talk describes advances of this paradigm along three interconnected thrusts: the design of triggering criteria that balance the trade-offs among performance, efficiency, and implementability; the synthesis of distributed triggers in network systems that can be evaluated by individual agents; and the benefits of flexibly interpreting what constitutes a resource. Throughout the presentation, we illustrate our discussion with applications to the coordination of human-robot interactions, the opportunistic actuation of safety-critical systems, information exchanges between neighboring agents in network systems, and accelerated optimization.

Software is disrupting one industry after the other. Currently, the automotive industry is feeling the pinch to innovate in the software business. New, innovative approaches to vehicles are needed. While many others connect single vehicles developed in isolation to build ecosystems, Bosch Research starts by reimagining the “fabric” connecting the world of AIoT, in which vehicles are one class of elements. The result is a disruptive technology we call Reliable Distributed Systems (RDS), enabling the operation of vehicles, where parts, such as sensing and compute are no longer bound to the vehicle, but can be performed inthe fabric.

This talk will address some key decisional issues that are necessary for a cognitive and interactive robot which shares space and tasks with humans. We adopt a constructive approach based on the identification and the effective implementation of individual and collaborative skills. The approach is comprehensive since it aims at dealing with a complete set of abilities articulated so that the robot controller is effectively able to conduct in a flexible and fluent manner a human-robot joint action seen as a collaborative problem solving and task achievement. These abilities include geometric reasoning and situation assessment based essentially on perspective-taking and affordances, management and exploitation of each agent (human and robot) knowledge in a separate cognitive model, human-aware task planning and interleaved execution of shared plans. We will also discuss the key issues linked to the pertinence and the acceptability by the human of the robot behaviour, and how this influence qualitatively the robot decisional, planning, control and communication processes.

There is a high possibility of communication system failure, packet loss, or hacking in a multi-agent system. To mitigate the relevant problems, we demonstrate swarm aggregation without any inter-agent communication. The solution is motivated by the movement of a school of fishes attacking food particles. The talk discusses a decentralized controller for aggregating a swarm of unicycle agents where each agent uses local sensing and avoids exchanging information. The methodology implicitly facilitates tasks like obstacle avoidance and path-following, whereas the agents exercise local decision-making. Using the theory of switched systems, we establish asymptotic stability of the formed aggregates. The proposed technique without communication demonstrates similar performance compared with existing swarm aggregation techniques using global and local communications.

This talk focuses on 3 CPS problems across autonomous vehicles, medical devices and energy systems. We will understand how modeling for such CPS problems requires a combination of formal methods, controls, and machine learning. Approaches to solving such problems highlight fundamental challenges in guarantees of safety and performance in data-driven CPS.
Theme 1 – Safe Autonomy: A Driver’s License Test for Driverless Vehicles Autonomous vehicles (AVs) have already driven millions of miles on public roads, but even the simplest maneuvers such as a lane change or vehicle overtake have not been certified for safety. To capture the long tail of safety cases we describe the design of a search engine for AV crashes.
Theme 2 – Safe Medical Devices: Computer-aided Clinical Trials Clinical trials can cost $10-20 million, last anywhere from 4-6 years and over 35% fail. We investigate how computer models and simulations of the physiology and medical devices in the closed-loop conduct in-silico trials and can be used as regulatory-grade evidence.
Theme 3 – Energy Systems: Learning and Control using Gaussian Processes Electricity markets have become increasingly volatile where 20-40X price spikes have become the norm. We explore data-driven approaches that bridge machine learning and controls for volatile energy markets.

Robotic agents that can accomplish manipulation tasks with the competence of humans have been one of the grand research challenges for artificial intelligence (AI) and robotics research for more than 50 years. However, while the fields made huge progress over the years, this ultimate goal is still out of reach. I believe that this is the case because the knowledge representation and reasoning methods that have been proposed in AI so far are necessary but too abstract. In this talk I propose to address this problem by endowing robots with the capability to internally emulate and simulate their perception-action loops based on realistic images and faithful physics simulations, which are made machine-understandable by casting them as virtual symbolic knowledge bases. These capabilities allow robots to generate huge collections of machine-understandable manipulation experiences, which robotic agents can generalize into commonsense and intuitive physics knowledge applicable to open varieties of manipulation tasks. The combination of learning, representation, and reasoning will equip robots with an understanding of the relation between their motions and the physical effects they cause at an unprecedented level of realism, depth, and breadth, and enable them to master human-scale manipulation tasks. This breakthrough will be achievable by combining leading- edge simulation and visual rendering technologies with mechanisms to semantically interpret and introspect internal simulation data structures and processes. With this talk we will give an outlook of applying CRAM to underwater missions

Artificial intelligence is leading advances in automating various aspects of science and medicine. With recent advances in automatically estimating whole-body human pose from videos, neuromotor rehabilitation is the next frontier for artificial intelligence in medicine. In this talk, I will outline our research on using neural network models for automated neuromotor disorder detection in infants. Infants have complex aperiodic movements that make it difficult, expensive, and time-intensive to quantify what ‘typically developing’ movements look like using traditional methods. To solve this, we curated a large dataset of typical infant movements using computer vision on videos, against which to assess neuromotor development. Using unsupervised learning methods, we estimated the probability that a given infant’s movements are typical as a measure of neuromotor disorder risk. In addition to body movements, we also created a system to automatically classify infant emotion from facial cues, which are an important behavioral signal of development. This system lays the foundations for accessible and quantified automated infant neuro-developmental assessment.

TBD