We propose a policy search approach to learn controllers from specifications given as Signal Temporal Logic (STL) formulae. The system model, which is unknown but assumed to be an affine control system, is learned together with the control policy. The model is implemented as two feedforward neural networks (FNNs) - one for the drift, and one for the control directions. To capture the history dependency of STL specifications, we use a recurrent neural network (RNN) to implement the control policy. In contrast to prevalent model-free methods, the learning approach proposed here takes advantage of the learned model and is more efficient. We use control barrier functions (CBFs) with the learned model to improve the safety of the system. We validate our algorithm via simulations and experiments. The results show that our approach can satisfy the given specification within very few system runs, and can be used for on-line control.
We propose a framework based on Recurrent Neural Networks (RNNs) to determine an optimal control strategy for a discrete-time system that is required to satisfy specifications given as Signal Temporal Logic (STL) formulae. RNNs can store information of a system over time, thus, enable us to determine satisfaction of the dynamic temporal requirements specified in STL formulae. Given a STL formula, a dataset of satisfying system executions and corresponding control policies, we can use RNNs to predict a control policy at each time based on the current and previous states of system. We use Contro…
In this paper, a method for learning a recurrent neural network (RNN) controller that maximizes the robustness of signal temporal logic (STL) specifications is presented. In contrast to previous methods, we consider synthesizing the RNN controller for which the user is able to select an STL specification arbitrarily from multiple STL specifications. To obtain such a controller, we propose a novel notion called STL2vec, which represents a vector representation of the STL specifications and exhibits their similarities. The construction of the STL2vec is useful since it allows us to enhance the e…
This paper introduces the safety controller architecture as a runtime assurance mechanism for system specifications expressed as safety properties in Linear Temporal Logic (LTL). The safety controller has three fundamental components: a performance controller, a backup controller, and an assurance mechanism. The assurance mechanism uses a monitor, constructed as a finite state machine (FSM), to analyze a suggested performance control input and search for system trajectories that are bad prefixes of the system specification. A fault flag from the assurance mechanism denotes a potentially danger…
Emerging applications in robotics and autonomous systems, such as autonomous driving and robotic surgery, often involve critical safety constraints that must be satisfied even when information about system models is limited. In this regard, we propose a model-free safety specification method that learns the maximal probability of safe operation by carefully combining probabilistic reachability analysis and safe reinforcement learning (RL). Our approach constructs a Lyapunov function with respect to a safe policy to restrain each policy improvement stage. As a result, it yields a sequence of sa…
We investigate a multi-agent planning problem, where each agent aims to achieve an individual task while avoiding collisions with others. We assume that each agent's task is expressed as a Time-Window Temporal Logic (TWTL) specification defined over a 3D environment. We propose a decentralized receding horizon algorithm for online planning of trajectories. We show that when the environment is sufficiently connected, the resulting agent trajectories are always safe (collision-free) and lead to the satisfaction of the TWTL specifications or their finite temporal relaxations. Accordingly, deadloc…
Stability and safety are two important aspects in safety-critical control of dynamical systems. It has been a well established fact in control theory that stability properties can be characterized by Lyapunov functions. Reachability properties can also be naturally captured by Lyapunov functions for finite-time stability. Motivated by safety-critical control applications, such as in autonomous systems and robotics, there has been a recent surge of interests in characterizing safety properties using barrier functions. Lyapunov and barrier functions conditions, however, are sometimes viewed as c…
As control systems become increasingly more complex, there exists a pressing need to find systematic ways of verifying them. To address this concern, there has been significant work in developing test generation schemes for black-box control architectures. These schemes test a black-box control architecture's ability to satisfy its control objectives, when these objectives are expressed as operational specifications through temporal logic formulae. Our work extends these prior, model based results by lower bounding the probability by which the black-box system will satisfy its operational spec…
Towards Safe Learning-Based Non-Linear Model Predictive Control through Recurrent Neural Network Modeling
Institute of Computer Engineering, TU Wien（计算机工程研究所，维也纳技术大学）；AIT Austrian Institute of Technology（奥地利技术研究所）；Institute of Mechanics and Mechatronics, TU Wien（机械与机电系统研究所，维也纳技术大学）
This paper proposes a safety analysis method that facilitates a tunable balance between the worst-case and risk-neutral perspectives. First, we define a risk-sensitive safe set to specify the degree of safety attained by a stochastic system. This set is defined as a sublevel set of the solution to an optimal control problem that is expressed using the Conditional Value-at-Risk (CVaR) measure. This problem does not satisfy Bellman's Principle, thus our next contribution is to show how risk-sensitive safe sets can be under-approximated by the solution to a CVaR-Markov Decision Process. We adopt …
With the advent of ISO 26262 there is an increased emphasis on top-down design in the automotive industry. ISO 26262 lacks detailed requirements for its various constituent phases. The lack of guidance becomes evident for the reuse of legacy components and subsystems, leaving vehicle architects and safety engineers to rely on experience without methodological support for their decisions. This poses challenges in the industry which is undergoing many significant changes due to new features like connectivity, electrification and automation. Here we focus on automated driving where multiple subsy…
In this paper, we use concepts from supervisory control theory of discrete event systems to propose a method to learn optimal control policies for a finite-state Markov Decision Process (MDP) in which (only) certain sequences of actions are deemed unsafe (respectively safe). We assume that the set of action sequences that are deemed unsafe and/or safe are given in terms of a finite-state automaton; and propose a supervisor that disables a subset of actions at every state of the MDP so that the constraints on action sequence are satisfied. Then we present a version of the Q-learning algorithm f…
Multi-agent partially observable Markov decision processes (MPOMDPs) provide a framework to represent heterogeneous autonomous agents subject to uncertainty and partial observation. In this paper, given a nominal policy provided by a human operator or a conventional planning method, we propose a technique based on barrier functions to design a minimally interfering safety-shield ensuring satisfaction of high-level specifications in terms of linear distribution temporal logic (LDTL). To this end, we use sufficient and necessary conditions for the invariance of a given set based on discrete-time…
We study the problem of refining satisfiability bounds for partially-known stochastic systems against planning specifications defined using syntactically co-safe Linear Temporal Logic (scLTL). We propose an abstraction-based approach that iteratively generates high-confidence Interval Markov Decision Process (IMDP) abstractions of the system from high-confidence bounds on the unknown component of the dynamics obtained via Gaussian process regression. In particular, we develop a synthesis strategy to sample the unknown dynamics by finding paths which avoid specification-violating states using a…
Smart cities are revolutionizing the transportation infrastructure by the integration of technology. However, ensuring that various transportation system components are operating as expected and in a safe manner is a great challenge. In this work, we propose the use of formal methods as a means to specify and reason about the traffic network's complex properties. Formal methods provide a flexible tool to define the safe operation of the traffic network by capturing non-conforming behavior, exploring various possible states of the traffic scene, and detecting any inconsistencies within it. Henc…
In this paper, we investigate joint sensor-actuator cyber attacks in discrete event systems. We assume that attackers can attack some sensors and actuators at the same time by altering observations and control commands. Because of the nondeterminism in observation and control caused by cyber attacks, the behavior of the supervised system becomes nondeterministic and may deviate from the safety specification. We define the upper-bound on all possible languages that can be generated by the supervised system to investigate the safety supervisory control problem under cyber attacks. After introduc…
Many safety-critical systems must achieve high-level task specifications with guaranteed safety and correctness. Much recent progress towards this goal has been made through controller synthesis from temporal logic specifications. Existing approaches, however, have been limited to relatively short and simple specifications. Furthermore, existing methods either consider some prior discretization of the state-space, deal only with a convex fragment of temporal logic, or are not provably complete. We propose a scalable, provably complete algorithm that synthesizes continuous trajectories to satis…
The prolific rise in autonomous systems has led to questions regarding their safe instantiation in real-world scenarios. Failures in safety-critical contexts such as human-robot interactions or even autonomous driving can ultimately lead to loss of life. In this context, this paper aims to provide a method by which one can algorithmically test and evaluate an autonomous system. Given a black-box autonomous system with some operational specifications, we construct a minimax problem based on control barrier functions to generate a family of test parameters designed to optimally evaluate whether …
In this work we formulate the satisfaction of a (syntactically co-safe) linear temporal logic specification on a physical plant through a recent hybrid dynamical systems formalism. In order to solve this problem, we introduce an extension to such a hybrid system framework of the so-called eventuality property, which matches suitably the condition for the satisfaction of such a temporal logic specification. The eventuality property can be established through barrier certificates, which we derive for the considered hybrid system framework. Using a hybrid barrier certificate, we propose a solutio…
