Back-Propagation Through Signal Temporal Logic Specifications: Infusing Logical Structure into Gradient-Based Methods

Leung, Karen; Aréchiga, Nikos; Pavone, Marco

doi:10.1007/978-3-030-66723-8_26

Cited by 33 publications

(52 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The robustness for space-time signals can be computed from a recursive definition (Donzé & Maler, 2010;G. E. Fainekos & Pappas, 2009;Leung et al, 2020;H. Yang, 2013).…”

Section: Cost Functionsmentioning

confidence: 99%

“…Yang, 2013). The derivative of the robustness with respect to the state trajectory can be approximated, which may help derive gradient-driven coverage metrics (Leung et al, 2020;Pant et al, 2017) or with doing gradient-based optimization if the black-box assumption is relaxed. Causal information in the form of a Bayesian network can be incorporated into the robustness for improved falsification (Akazaki et al, 2017).…”

Section: Cost Functionsmentioning

confidence: 99%

Section: Cost Functionsmentioning

confidence: 99%

“…Robustness can be a challenging objective to optimize because it can be non-smooth (Leung et al, 2020) and be dominated by state variables with larger magnitude values. Leung et al (2020) propose smooth approximations to the robustness and Akazaki et al (2018) optimize a convex function of the robustness −exp (−ρ(s)), which bounds the maximum cost. State variables may be normalized to alleviate differences in scale between the variables, but normalization requires the range of the state variables which may not be known a priori.…”

Section: Cost Functionsmentioning

confidence: 99%

See 3 more Smart Citations

A Survey of Algorithms for Black-Box Safety Validation of Cyber-Physical Systems

CorsoAnthony¹,

Moss

KorenMark

et al. 2021

jair

View full text Add to dashboard Cite

Autonomous cyber-physical systems (CPS) can improve safety and efficiency for safety-critical applications, but require rigorous testing before deployment. The complexity of these systems often precludes the use of formal verification and real-world testing can be too dangerous during development. Therefore, simulation-based techniques have been developed that treat the system under test as a black box operating in a simulated environment. Safety validation tasks include finding disturbances in the environment that cause the system to fail (falsification), finding the most-likely failure, and estimating the probability that the system fails. Motivated by the prevalence of safety-critical artificial intelligence, this work provides a survey of state-of-the-art safety validation techniques for CPS with a focus on applied algorithms and their modifications for the safety validation problem. We present and discuss algorithms in the domains of optimization, path planning, reinforcement learning, and importance sampling. Problem decomposition techniques are presented to help scale algorithms to large state spaces, which are common for CPS. A brief overview of safety-critical applications is given, including autonomous vehicles and aircraft collision avoidance systems. Finally, we present a survey of existing academic and commercially available safety validation tools.

show abstract

“…The robustness for space-time signals can be computed from a recursive definition (Donzé & Maler, 2010;G. E. Fainekos & Pappas, 2009;Leung et al, 2020;H. Yang, 2013).…”

Section: Cost Functionsmentioning

confidence: 99%

Section: Cost Functionsmentioning

confidence: 99%

Section: Cost Functionsmentioning

confidence: 99%

Section: Cost Functionsmentioning

confidence: 99%

See 2 more Smart Citations

A Survey of Algorithms for Black-Box Safety Validation of Cyber-Physical Systems

CorsoAnthony¹,

Moss

KorenMark

et al. 2021

jair

View full text Add to dashboard Cite

show abstract

“…We evaluate the proposed method on 8 benchmark synthesis problems with a variety of different scenarios. We compare with both abstraction-based and MPC-based methods [5], [6]. Empirical results show that our method outperforms other state-of-the-art methods in terms of running time and quality of the planned paths, not to say that our method can handle much more general STL formulas.…”

Section: Introductionmentioning

confidence: 99%

Multi-Agent Motion Planning From Signal Temporal Logic Specifications

Sun

Chen

Mitra

et al. 2022

IEEE Robot. Autom. Lett.

View full text Add to dashboard Cite

We tackle the challenging problem of multi-agent cooperative motion planning for complex tasks described using signal temporal logic (STL), where robots can have nonlinear and nonholonomic dynamics. Existing methods in multi-agent motion planning, especially those based on discrete abstractions and model predictive control (MPC), suffer from limited scalability with respect to the complexity of the task, the size of the workspace, and the planning horizon. We present a method based on timed waypoints to address this issue. We show that timed waypoints can help abstract nonlinear behaviors of the system as safety envelopes around the reference path defined by those waypoints. Then the search for waypoints satisfying the STL specifications can be inductively encoded as a mixedinteger linear program. The agents following the synthesized timed waypoints have their tasks automatically allocated, and are guaranteed to satisfy the STL specifications while avoiding collisions. We evaluate the algorithm on a wide variety of benchmarks. Results show that it supports multi-agent planning from complex specification over long planning horizons, and significantly outperforms state-of-the-art abstraction-based and MPC-based motion planning methods. The implementation is available at https://github.com/sundw2014/STLPlanning.

show abstract

Classification of Driving Behaviors Using STL Formulas: A Comparative Study

Karagulle

Aréchiga²,

DeCastro³

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

In this paper, we conduct a preliminary comparative study of classification of longitudinal driving behavior using Signal Temporal Logic (STL) formulas. The goal of the classification problem is to distinguish between different driving styles or vehicles. The results can be used to design and test autonomous vehicle policies. We work on a reallife dataset, the Highway Drone Dataset (HighD). To solve this problem, our first approach starts with a formula template and reduces the classification problem to a Mixed-Integer Linear Program (MILP). Solving MILPs becomes computationally challenging with increasing number of variables and constraints. We propose two improvements to split the classification problem into smaller ones. We prove that these simpler problems are related to the original classification problem in a way that their feasibility imply that of the original. Finally, we compare our MILP formulation with an existing STL-based classification tool, LoTuS, in terms of accuracy and execution time.

show abstract

Back-Propagation Through Signal Temporal Logic Specifications: Infusing Logical Structure into Gradient-Based Methods

Cited by 33 publications

References 16 publications

A Survey of Algorithms for Black-Box Safety Validation of Cyber-Physical Systems

A Survey of Algorithms for Black-Box Safety Validation of Cyber-Physical Systems

Multi-Agent Motion Planning From Signal Temporal Logic Specifications

Classification of Driving Behaviors Using STL Formulas: A Comparative Study

Contact Info

Product

Resources

About