In a neural network with ReLU activations, the number of piecewise linear regions in the output can grow exponentially with depth. However, this is highly unlikely to happen when the initial parameters are sampled randomly, which therefore often leads to the use of networks that are unnecessarily large. To address this problem, we introduce a novel parameterization of the network that restricts its weights so that a depth $d$ network produces exactly $2^d$ linear regions at initialization and maintains those regions throughout training under the parameterization. This approach allows us to learn approximations of convex, one-dimensional functions that are several orders of magnitude more accurate than their randomly initialized counterparts. We further demonstrate a preliminary extension of our construction to multidimensional and non-convex functions, allowing the technique to replace traditional dense layers in various architectures.

Two-vehicle racing is natural example of a competitive dynamic game. As with most dynamic games, there are many ways in which the underlying solution concept can be structured, resulting in different equilibrium concepts. The assumed solution concept influences the behaviors of two interacting players in racing. For example, blocking behavior emerges naturally in leader-follower play, but to achieve this in Nash play the costs would have to be chosen specifically to trigger this behavior. In this work, we develop a novel model for competitive two-player vehicle racing, represented as an equilibrium problem, complete with simplified aerodynamic drag and drafting effects, as well as position-dependent collisionavoidance responsibility. We use our model to explore how different solution concepts affect competitiveness. We develop a solution for bilevel optimization problems, enabling a largescale empirical study comparing bilevel strategies (either as leader or follower), Nash equilibrium strategy and a singleplayer constant velocity baseline. We find the choice of strategies significantly affects competitive performance and safety.

Collision detection is a critical functionality for robotics. The degree to which objects collide cannot be represented as a continuously differentiable function for any shapes other than spheres. This paper proposes a framework for handling collision detection between polyhedral shapes. We frame the signed distance between two polyhedral bodies as the optimal value of a convex optimization, and consider constraining the signed distance in a bilevel optimization problem. To avoid relying on specialized bilevel solvers, our method exploits the fact that the signed distance is the minimal point of a convex region related to the two bodies. Our method enumerates the values obtained at all extreme points of this region and lists them as constraints in the higher-level problem. We compare our formulation to existing methods in terms of reliability and speed when solved using the same mixed complementarity problem solver. We demonstrate that our approach more reliably solves difficult collision detection problems with multiple obstacles than other methods, and is faster than existing methods in some cases.

Integer programming games (IPGs) are a popular game-theoretic tool to model an array of games where each player has a discrete strategy set. These games arise in important domains such as economics, transportation, cybersecurity, etc., but solving them is non-trivial as it is known that checking for the existence of pure Nash equilibria in an IPG is Σ₂^p-complete. Recent works have proposed a class of relaxed solution concepts for IPGs called locally optimal integer solutions (LOIS) and shown it to be an efficient alternative for pure Nash equilibria. While LOIS are significantly simpler to compute, they still do not scale when solved using traditional mathematical solvers, especially when high-quality solutions are desired. In this paper, we apply commercially available SAT solvers to find LOIS in IPGs. We investigate efficient encodings for a cybersecurity game and compare solution times when using SAT solvers vs mathematical program solvers. We also investigate the application of SAT solvers in graph games using a graph interdiction example and compare against the obtained LOI solutions against existing heuristics-based solutions. Our results indicate that with appropriate encodings, large-scale IPGs can be solved much more efficiently using SAT solvers. We also show that SAT solvers can be applied to graph games in conjunction with LOIS for obtaining high-quality solutions. Our results emphasize the potential of SAT solvers combined with LOIS to solve significant game theory problems.

As automation increases qualitatively and quantitatively in safety-critical human cyber-physical systems, it is becoming more and more challenging to increase the probability or ensure that human operators still perceive key artifacts and comprehend their roles in the system. In the companion paper, we proposed an abstract reference architecture capable of expressing all classes of system-level interactions in human cyber-physical systems. Here we demonstrate how this reference architecture supports the analysis of levels of communication between agents and helps to identify the potential for misunderstandings and misconceptions. We then develop a metamodel for safe human machine interaction. Therefore, we ask what type of information exchange must be supported on what level so that humans and systems can cooperate as a team, what is the criticality of exchanged information, what are timing requirements for such interactions, and how can we communicate highly critical information in a limited time frame in spite of the many sources of a distorted perception. We highlight shared stumbling blocks and illustrate shared design principles, which rest on established ontologies specific to particular application classes. In order to overcome the partial opacity of internal states of agents, we anticipate a key role of virtual twins of both human and technical cooperation partners for designing a suitable communication.

The design and analysis of multi-agent human cyber-physical systems in safety-critical or industry-critical domains calls for an adequate semantic foundation capable of exhaustively and rigorously describing all emergent effects in the joint dynamic behavior of the agents that are relevant to their safety and well-behavior. We present such a semantic foundation. This framework extends beyond previous approaches by extending the agent-local dynamic state beyond state components under direct control of the agent and belief about other agents (as previously suggested for understanding cooperative as well as rational behavior) to agent-local evidence and belief about the overall cooperative, competitive, or coopetitive game structure. We argue that this extension is necessary for rigorously analyzing systems of human cyber-physical systems because humans are known to employ cognitive replacement models of system dynamics that are both non-stationary and potentially incongruent. These replacement models induce visible and potentially harmful effects on their joint emergent behavior and the interaction with cyber-physical system components.

We present a novel method for handling uncertainty about the intentions of non-ego players in trajectory games, with application to motion planning for autonomous vehicles. Our method models the uncertainty about the intention of other agents by constructing multiple hypotheses about the objectives and constraints of other agents in the scene. For each candidate hypothesis, we associate a Bernoulli random variable representing the probability of that hypothesis, which may or may not be independent of the probability of other hypotheses. We leverage constraint asymmetries and feedback information patterns to incorporate the probabilities of hypotheses in a natural way. Specifically, increasing the probability associated with a given hypothesis from 0 to 1 shifts the responsibility of collision avoidance from the hypothesized agent to the ego agent. This method allows the generation of interactive trajectories for the ego agent, where the level of assertiveness or caution that the ego exhibits is directly related to the easy-to-model uncertainty it maintains about the scene.

A method is presented for solving the discrete-time finite-horizon Linear Quadratic Regulator (LQR) problem subject to auxiliary linear equality constraints, such as fixed end-point constraints. The method explicitly determines an affine relationship between the control and state variables, as in standard Riccati recursion, giving rise to feedback control policies that account for constraints. Since the linearly-constrained LQR problem arises commonly in robotic trajectory optimization, having a method that can efficiently compute these solutions is important. We demonstrate some of the useful properties and interpretations of said control policies, and we compare the computation time and complexity of our method against existing methods.