When autonomous robots begin to share the human living and working spaces, safety becomes paramount. It is legally required that the safety of such systems is ensured, e.g. by certification according to relevant standards such as IEC 61508. However, such safety considerations are usually not addressed in academic robotics. In this paper we report on one such successful endeavor, which is concerned with designing, implementing, and certifying a collision avoidance safety function for autonomous vehicles and static obstacles. The safety function calculates a safety zone for the vehicle, depending on its current motion, which Electronic supplementary material The online version of this article (is as large as required but as small as feasible, thus ensuring safety against collision with static obstacles. We outline the algorithm which was specifically designed with safety in mind, and present our verification methodology which is based on formal proof and verification using the theorem prover Isabelle. The implementation and our methodology have been certified for use in applications up to SIL 3 of IEC 61508 by a certification authority (TÜV Süd Rail GmbH, Germany). Throughout, issues we recognized as being important for a successful application of formal methods in robotics are highlighted. Moreover, we argue that formal analysis deepens the understanding of the algorithm, and hence is valuable even outside the safety context.

No abstract

This paper introduces the processing element architecture of the second generation SpiNNaker chip, implemented in 22nm FDSOI. On circuit level, the chip features adaptive body biasing for near-threshold operation, and dynamic voltage-andfrequency scaling driven by spiking activity. On system level, processing is centered around an ARM M4 core, similar to the processor-centric architecture of the first generation SpiNNaker. To speed operation of subtasks, we have added accelerators for numerical operations of both spiking (SNN) and rate based (deep) neural networks (DNN). PEs communicate via a dedicated, custom-designed network-on-chip. We present three benchmarks showing operation of the whole processor element on SNN, DNN and hybrid SNN/DNN networks.

No abstract

Language-based securityInformation-flow analysis Dynamic logic Security type system Formal verification a b s t r a c t Type systems and program logics are often thought to be at opposing ends of the spectrum of formal software analyses. In this paper we show that a flow-sensitive type system ensuring non-interference in a simple while-language can be expressed through specialised rules of a program logic. In our framework, the structure of non-interference proofs resembles the corresponding derivations in a state-of-the-art security type system, meaning that the algorithmic version of the type system can be used as a proof procedure for the logic. We argue that this is important for obtaining uniform proof certificates in a proof-carrying code framework. We discuss in which cases the interleaving of approximative and precise reasoning allows us to deal with delimited information release. Finally, we present ideas on how our results can be extended to encompass features of realistic programming languages such as Java.