An Intelligent Task Scheduling Mechanism for Autonomous Vehicles via Deep Learning

Abstract

With the rapid development of the Internet of Things (IoT) and artificial intelligence, autonomous vehicles have received much attention in recent years. Safe driving is one of the essential concerns of self-driving cars. The main problem in providing better safe driving requires an efficient inference system for real-time task management and autonomous control. Due to limited battery life and computing power, reducing execution time and resource consumption can be a daunting process. This paper addressed these challenges and developed an intelligent task management system for IoT-based autonomous vehicles. For each task processing, a supervised resource predictor is invoked for optimal hardware cluster selection. Tasks are executed based on the earliest hyper period first (EHF) scheduler to achieve optimal task error rate and schedule length performance. The single-layer feedforward neural network (SLFN) and lightweight learning approaches are designed to distribute each task to the appropriate processor based on their emergency and CPU utilization. We developed this intelligent task management module in python and experimentally tested it on multicore SoCs (Odroid Xu4 and NVIDIA Jetson embedded platforms).Connected Autonomous Vehicles (CAV) and Internet of Medical Things (IoMT) benchmarks are used for training and testing purposes. The proposed modules are validated by observing the task miss rate, resource utilization, and energy consumption metrics compared with state-of-art heuristics. SLFN-EHF task scheduler achieved better results in an average of 98% accuracy, and in an average of 20–27% reduced in execution time and 32–45% in task miss rate metric than conventional methods.

Evaluation of Thermal Imaging on Embedded GPU Platforms for Application in Vehicular Assistance Systems

“…In another study [38] authors proposed deep learning-based smart task scheduling for self-driving vehicles. This task management module was implemented on multicore SoCs (Odroid Xu4 and Nvidia Jetson).…”

Deep Learning and Autonomous Vehicles: Strategic Themes, Applications, and Research Agenda Using SciMAT and Content-Centric Analysis, a Systematic Review

“…• Energy: Vega et al [59] and Balasekaran et al [60] highlighted the issue of energy efficiency, owing to the fact that AI models, in general, have a high computational cost and, as a result, require large amounts of energy resources; • Legislation: Thiele-Evans et al [61] and Ning et al [62] discussed the legislation of commercial implementation of AVs, comparing the progress in this area in various countries; • Ethics: Cunneen et al [63] addressed ethics in AV projects, a topic that is directly related to legislation, because in many countries, the main debate for approving commercial use of AVs is about ethical barriers. The trolley problem is the most wellknown ethical issue concerning AVs [60]; • Cybersecurity: Khan et al [64] and Deng et al [48] concentrated on AI models that can aid in the detection of AV computational system attacks. Furthermore, Jiang et al [65] investigated how attacks on AVs are carried out by experimentally replicating Other datasets identified, such as CityScapes, BDD100K, ApolloScape, Caltech, CamVId, nuScenes, PASCAL, and Waymo Open Dataset, are very similar to KITTI but were created by other research groups from universities or companies, using different platforms and sensors for data collection.…”

DDDQN‐TS: A task scheduling and load balancing method based on optimized deep reinforcement learning in heterogeneous computing environment

“…In recent years with the rapid development of artificial intelligence technology, machine learning methods are gradually being applied to task scheduling problems. [6][7][8]20,21 Xie et al 6 employed MultiLogistic Regression theory (called MLRS for short) on task scheduling issue in HCE, the model of which is trained using the data collected from historical best scheduling plans. Besides research work using traditional machine learning methods, there are also some studies done using neural networks.…”

“…When the training episode reaches , the Q-network begins learning from past experiences. In this stage (lines [17][18][19][20][21][22][23][24], in each episode a minibatch of size k will be randomly sampled from the replay buffer, and target Q′ will be calculated based on Double Q-learning method while estimated Q will be predicted by the Q-Network. Then the MSE can be calculated (line 20).…”