Guidance systems for visually impaired persons have become a popular topic in recent years. Existing guidance systems on the market typically utilize auxiliary tools and methods such as GPS, UWB, or a simple white cane that exploits the user’s single tactile or auditory sense. These guidance methodologies can be inadequate in a complex indoor environment. This paper proposes a multi-sensory guidance system for the visually impaired that can provide tactile and auditory advice using ORB-SLAM and YOLO techniques. Based on an RGB-D camera, the local obstacle avoidance system is realized at the tactile level through point cloud filtering that can inform the user via a vibrating motor. Our proposed method can generate a dense navigation map to implement global obstacle avoidance and path planning for the user through the coordinate transformation. Real-time target detection and a voice-prompt system based on YOLO are also incorporated at the auditory level. We implemented the proposed system as a smart cane. Experiments are performed using four different test scenarios. Experimental results demonstrate that the impediments in the walking path can be reliably located and classified in real-time. Our proposed system can function as a capable auxiliary to help visually impaired people navigate securely by integrating YOLO with ORB-SLAM.
Health and safety management has been an important issue in construction industry. National regulations impose the using of hard hats in construction sites. However, there are often cases in which the construction workers neglect the regulations. It is desired to monitor the correct wearing of hard hat in real time and explore monitoring techniques facilitated by deep-learning algorithms. In this paper, a convolutional neural network based hard-hat detection algorithm is proposed. In this algorithm, the detection of construction workers and the hard hats are assisted by computer vision technique where deep learning model are trained to identify the proper wearing of hard hats. The optimization of the proposed neural networks can reduce the computational complexity while maintaining a relatively high recognition precision. Experiments have been performed using five different algorithms for comparison and results demonstrate that the proposed algorithm excels in the mAP and FPS performance metrics. The experimental results collected on an embedded platform reveal that the proposed algorithm presents a good candidate for similar applications where real-time deep-learning application is desired.
20 October 2019 in International Journal of Modern Physics A
The convergence of artificial intelligence and the Internet of Things (IoT) has made remarkable strides in the realm of industry. In the context of AIoT edge computing, where IoT devices collect data from diverse sources and send them for real-time processing at edge servers, existing message queue systems face challenges in adapting to changing system conditions, such as fluctuations in the number of devices, message size, and frequency. This necessitates the development of an approach that can effectively decouple message processing and handle workload variations in the AIoT computing environment. This study presents a distributed message system for AIoT edge computing, specifically designed to address the challenges associated with message ordering in such environments. The system incorporates a novel partition selection algorithm (PSA) to ensure message order, balance the load among broker clusters, and enhance the availability of subscribable messages from AIoT edge devices. Furthermore, this study proposes the distributed message system configuration optimization algorithm (DMSCO), based on DDPG, to optimize the performance of the distributed message system. Experimental evaluations demonstrate that, compared to the genetic algorithm and random searching, the DMSCO algorithm can provide a significant improvement in system throughput to meet the specific demands of high-concurrency AIoT edge computing applications.
28 March 2024 in CAAI Transactions on Intelligence Technology
Multi‐agent reinforcement learning relies on reward signals to guide the policy networks of individual agents. However, in high‐dimensional continuous spaces, the non‐stationary environment can provide outdated experiences that hinder convergence, resulting in ineffective training performance for multi‐agent systems. To tackle this issue, a novel reinforcement learning scheme, Mutual Information Oriented Deep Skill Chaining (MioDSC), is proposed that generates an optimised cooperative policy by incorporating intrinsic rewards based on mutual information to improve exploration efficiency. These rewards encourage agents to diversify their learning process by engaging in actions that increase the mutual information between their actions and the environment state. In addition, MioDSC can generate cooperative policies using the options framework, allowing agents to learn and reuse complex action sequences and accelerating the convergence speed of multi‐agent learning. MioDSC was evaluated in the multi‐agent particle environment and the StarCraft multi‐agent challenge at varying difficulty levels. The experimental results demonstrate that MioDSC outperforms state‐of‐the‐art methods and is robust across various multi‐agent system tasks with high stability.
21 August 2025 in Expert Systems with Applications