© 2016 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works. Reproduced in accordance with the publisher's self-archiving policy.eprints@whiterose.ac.uk https://eprints.whiterose.ac.uk/ Reuse Unless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version -refer to the White Rose Research Online record for this item. Where records identify the publisher as the copyright holder, users can verify any specific terms of use on the publisher's website. TakedownIf you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request. Sampling-based Stochastic Optimal Control with Metric Interval Temporal Logic SpecificationsFelipe J. Montana, Jun Liu and Tony J. DoddAbstract-This paper describes a method to find optimal policies for stochastic dynamic systems that maximise the probability of satisfying real-time properties. The method consists of two phases. In the first phase, a coarse abstraction of the original system is created. In each region of the abstraction, a sampling-based algorithm is utilised to compute local policies that allow the system to move between regions. Then, in the second phase, the selection of a policy in each region is obtained by solving a reachability problem on the Cartesian product between the abstraction and a timed automaton representing a real-time specification given as a metric interval temporal logic formula. In contrast to current methods that require a fine abstraction, the proposed method achieves computational tractability by modelling the coarse abstraction of the system as a bounded-parameter Markov decision process (BMDP). Moreover, once the BMDP is created, this can be reused for new specifications assuming the same stochastic system and workspace. The method is demonstrated with an autonomous driving example.

Gas turbine engines are complex machines that typically generate a vast amount of data, and require careful monitoring to allow for cost-effective preventative maintenance. In aerospace applications, returning all measured data to ground is prohibitively expensive, often causing useful, high value, data to be discarded. The ability to detect, prioritise, and return useful data in real-time is therefore vital. This paper proposes that system output measurements, described by a convolutional neural network model of normality, are prioritised in real-time for the attention of preventative maintenance decision makers.Due to the complexity of gas turbine engine timevarying behaviours, deriving accurate physical models is difficult, and often leads to models with low prediction accuracy and incompatibility with real-time execution. Data-driven modelling is a desirable alternative producing high accuracy, asset specific models without the need for derivation from first principles.We present a data-driven system for online detection and prioritisation of anomalous data. Biased data assessment deriving from novel operating conditions is avoided by uncertainty management integrated into the deep neural predictive model. Testing is performed on real and synthetic data, showing sensitivity to both real and synthetic faults. The system is capable of running in real-time on low-power embedded hardware and is currently in deployment on the Rolls-Royce Pearl 15 engine flight trials.

ReuseUnless indicated otherwise, fulltext items are protected by copyright with all rights reserved. The copyright exception in section 29 of the Copyright, Designs and Patents Act 1988 allows the making of a single copy solely for the purpose of non-commercial research or private study within the limits of fair dealing. The publisher or other rights-holder may allow further reproduction and re-use of this version -refer to the White Rose Research Online record for this item. Where records identify the publisher as the copyright holder, users can verify any specific terms of use on the publisher's website. TakedownIf you consider content in White Rose Research Online to be in breach of UK law, please notify us by emailing eprints@whiterose.ac.uk including the URL of the record and the reason for the withdrawal request. Abstract. This paper presents a method that allows mobile systems with uncertainty in motion and sensing to react to unknown environments while high-level specifications are satisfied. Although previous works have addressed the problem of synthesising controllers under uncertainty constraints and temporal logic specifications, reaction to dynamic environments has not been considered under this scenario. The method uses feedback-based information roadmaps (FIRMs) to break the curse of history associated with partially observable systems. A transition system is incrementally constructed based on the idea of FIRMs by adding nodes on the belief space. Then, a policy is found in the product Markov decision process created between the transition system and a Rabin automaton representing a linear temporal logic formula. The proposed solution allows the system to react to previously unknown elements in the environment. To achieve fast reaction time, a FIRM considering the probability of violating the specification in each transition is used to drive the system towards local targets or to avoid obstacles. The method is demonstrated with an illustrative example.

This paper addresses the problem of path planning for multiple robots under high-level specifications given as syntactically co-safe linear temporal logic formulae. Most of the existing solutions use the notion of abstraction to obtain a discrete transition system that simulates the dynamics of the robot. Nevertheless, these solutions have poor scalability with the dimension of the configuration space of the robots. For problems with a single robot, sampling-based methods have been presented as a solution to alleviate this limitation. The proposed solution extends the idea of sampling methods to the multiple robot case. The method samples the configuration space of the robots to incrementally constructs a transition system that models the motion of all the robots as a group. This transition system is then combined with a Büchi automaton, representing the specification, in a Cartesian product. The product is updated with each expansion of the transition system until a solution is found. We also present a new algorithm that improves the performance of the proposed method by guiding the expansion of the transition system. The method is demonstrated with examples considering different number of robots and specifications.