Summary Reinforcement learning (RL) is an effective approach to developing control policies by maximizing the agent's reward. Deep reinforcement learning uses deep neural networks (DNNs) for function approximation in RL, and has achieved tremendous success in recent years. Large DNNs often incur significant memory size and computational overheads, which may impede their deployment into resource‐constrained embedded systems. For deployment of a trained RL agent on embedded systems, it is necessary to compress the policy network of the RL agent to improve its memory and computation efficiency. In this article, we perform model compression of the policy network of an RL agent by leveraging the relevance scores computed by layer‐wise relevance propagation (LRP), a technique for Explainable AI (XAI), to rank and prune the convolutional filters in the policy network, combined with fine‐tuning with policy distillation. Performance evaluation based on several Atari games indicates that our proposed approach is effective in reducing model size and inference time of RL agents. We also consider robust RL agents trained with RADIAL‐RL versus standard RL agents, and show that a robust RL agent can achieve better performance (higher average reward) after pruning than a standard RL agent for different attack strengths and pruning rates.

All systems rely on inputs and outputs (I/Os) to perceive and interact with their surroundings. In safety-critical systems, it is important to guarantee both the performance and time-predictability of I/O operations. However, with the continued growth of architectural complexity in modern safety-critical systems, satisfying such real-time requirements has become increasingly challenging due to complex I/O transaction paths and extensive hardware contention. In this article, we present a new Network-on-Chip (NoC)-based Predictable I/O system framework (NPRC-I/O) which reduces this contention and ensures the performance and time-predictability of I/O operations. Specifically, NPRC-I/O contains a programmable I/O command controller (NPRC-CC) and a run-time reconfigurable NoC ($\text{R}^{2}$NoC), which provides the capability to adjust I/O transaction paths at run time. Using this flexibility, we construct an end-to-end transmission latency analysis and an optimization engine that produces configurations for NPRC-I/O and the I/O traffic in a given system. The constructed analysis and optimization engine guarantee the timing of all hard real-time traffic while reducing the deadline misses of soft real-time traffic and overall transmission latency.

Reinforcement Learning (RL) is an effective approach to developing control policies by maximizing the agent’s reward. Deep Reinforcement Learning (DRL) uses Deep Neural Networks (DNNs) for function approximation in RL, and has achieved tremendous success in recent years. Large DNNs often incur significant memory size and computational overheads, which greatly impedes their deployment into resource-constrained embedded systems. For deployment of a trained RL agent on embedded systems, it is necessary to compress the Policy Network of the RL agent to improve its memory and computation efficiency. In this paper, we perform model compression of the Policy Network of an RL agent by leveraging the relevance scores computed by Layer-wise Relevance Propagation (LRP), a technique for Explainable AI (XAI), to rank and prune the convolutional filters in the Policy Network, combined with fine-tuning with Policy Distillation. Performance evaluation based on several Atari games indicates that our proposed approach is effective in reducing model size and inference time of RL agents.

The Controller Area Network (CAN) is a ubiquitous bus protocol present in the Electrical/Electronic (E/E) systems of almost all vehicles. It is vulnerable to a range of attacks once the attacker gains access to the bus through the vehicle’s attack surface. We address the problem of Intrusion Detection on the CAN bus and present a series of methods based on two classifiers trained with Auxiliary Classifier Generative Adversarial Network (ACGAN) to detect and assign fine-grained labels to Known Attacks and also detect the Unknown Attack class in a dataset containing a mixture of (Normal + Known Attacks + Unknown Attack) messages. The most effective method is a cascaded two-stage classification architecture, with the multi-class Auxiliary Classifier in the first stage for classification of Normal and Known Attacks, passing Out-of-Distribution (OOD) samples to the binary Real-Fake Classifier in the second stage for detection of the Unknown Attack class. Performance evaluation demonstrates that our method achieves both high classification accuracy and low runtime overhead, making it suitable for deployment in the resource-constrained in-vehicle environment.

A Mixed-Criticality System (MCS) features the integration of multiple subsystems that are subject to different levels of safety certification on a shared hardware platform. In cost-sensitive application domains such as automotive E/E systems, it is important to reduce application memory footprint, since such a reduction may enable the adoption of a cheaper microprocessor in the family. Preemption Threshold Scheduling (PTS) is a well-known technique for reducing system stack usage. We consider partitioned multiprocessor scheduling, with Preemption Threshold Adaptive Mixed-Criticality (PT-AMC) as the task scheduling algorithm on each processor and address the optimization problem of finding a feasible task-to-processor mapping with minimum total system stack usage on a resource-constrained multi-processor. We present the Extended Maximal Preemption Threshold Assignment Algorithm (EMPTAA), with dual purposes of improving the taskset’s schedulability if it is not already schedulable, and minimizing system stack usage of the schedulable taskset. We present efficient heuristic algorithms for finding sub-optimal yet high-quality solutions, including Maximum Utilization Difference based Partitioning (MUDP) and MUDP with Backtrack Mapping (MUDP-BM), as well as a Branch-and-Bound (BnB) algorithm for finding the optimal solution. Performance evaluation with synthetic task sets demonstrates the effectiveness and efficiency of the proposed algorithms.

Model-based design using Synchronous Reactive (SR) models enables early design and verification of application functionality in a platform-independent manner, and the implementation on the target platform should guarantee the preservation of application semantic properties. Mixed-Criticality Scheduling (MCS) is an effective approach to addressing diverse certification requirements of safety-critical systems that integrate multiple subsystems with different levels of criticality. This article considers fixed-priority scheduling of mixed-criticality SR models, and considers two scheduling approaches: Adaptive MCS and Elastic MCS. We formulate the optimization problem of minimizing the total system cost of added functional delays in the implementation while guaranteeing schedulability, and present an optimal algorithm based on branch-and-bound search, and an efficient heuristic algorithm.

Automotive in-vehicle systems are distributed systems consisting of multiple ECUs (Electronic Control Units) interconnected with a broadcast network such as FlexRay. Message authentication is an effective mechanism to prevent attackers from injecting malicious messages into the network. In order to reduce timing interference of message authentication operations on application tasks, hardware coprocessors in the form of either FPGA or ASIC are adopted to offload computation-intensive cryptographic algorithms from the ECU. However, it may not be feasible or desirable to equip every ECU with a hardware coprocessor, as modern vehicles can contain more than one hundred ECUs, and the automotive industry is cost-sensitive. In this paper, we consider the problem of mapping an application task graph onto a FlexRay-based distributed hardware platform, to meet security and deadline requirements while minimizing the number of hardware coprocessors needed in the system. We present a Mixed Integer Linear Programming (MILP) formulation, a divide-and-conquer heuristic algorithm, and a Simulated Annealing algorithm. We evaluate the algorithms with industrial case studies.

Mixed Criticality Systems (MCS) have gained increasing interest in the past few years due to their industrial relevance. When mixed-criticality systems are implemented on multicore architectures, several challenges arise such as the efficient partitioning of these systems. In this paper, we address this issue by presenting a novel mixed-criticality partitioning algorithm, the Dual-Partitioned Mixed-Criticality (DPM) algorithm, that allows limited migration of LO-criticality tasks to enhance the efficiency of the partitioning while maintaining many of the advantages of partitioned systems. Experimental results show that DPM consistently outperforms existing mixed-criticality partitioning algorithms, for example, at utilizations of 0.8 or higher, DPM is able to schedule 17% more systems.

In a mixed-criticality system, multiple tasks with different levels of criticality may coexist on the same hardware platform. The scheduling algorithm EDF-VD (Earliest Deadline First with Virtual Deadlines) has been proposed for mixed-criticality systems, which assumes tasks do not share any common resources. We present MC-SRP (Mixed-Criticality Stack Resource Policy), a resource synchronization protocol for EDF-VD, which allows resource sharing among tasks at the same criticality level and guarantees that each task is blocked at most once in each criticality mode. In addition, we present MC-SRPT (MC-SRP with Thresholds) for reducing the application stack size requirement in resource-constrained embedded systems.

Mixed-Criticality Scheduling (MCS) is an effective approach to addressing diverse certification requirements of safety-critical systems that integrate multiple subsystems with different levels of criticality. Preemption Threshold Scheduling (PTS) is a well-known technique for controlling the degree of preemption, ranging from fully-preemptive to fully-non-preemptive scheduling. We present schedulability analysis algorithms to enable integration of PTS with MCS, in order to bring the rich benefits of PTS into MCS, including minimizing the application stack space requirement, reducing the number of runtime task preemptions, and improving schedulability.

In mixed-criticality systems, multiple subsystems with different levels of criticality may co-exist on the same hardware platform. Many scheduling algorithms have been proposed to achieve certification at multiple levels of criticality. However, current MCS algorithms and analysis techniques generally assume tasks are independent, i.e., they do not share data that need to be protected with synchronization mechanisms like mutexes or semaphores. In this paper, we address this limitation by presenting an extension to the Stack Resource Protocol (SRP), called Mixed-Criticality-SRP (MC-SRP). Moreover, preemption-threshold scheduling is a well-known technique for reducing stack space size and enhance schedulability in resource-constrained embedded systems.We also present the integration of preemption-thresholds into EDF-based mixed-criticality scheduling (MCS) algorithms, and develop the schedulability analysis methods to such systems.