Navigation is a crucial challenge for mobile robots. Currently, deep reinforcement learning has attracted considerable attention and has witnessed substantial development owing to its robust performance and learning capabilities in real-world scenarios. Scientists leverage the advantages of deep neural networks, such as long short-term memory, recurrent neural networks, and convolutional neural networks, to integrate them into mobile robot navigation based on deep reinforcement learning. This integration aims to enhance the robot's motion control performance in both static and dynamic environments. This paper illustrates a comprehensive survey of deep reinforcement learning methods applied to mobile robot navigation systems in crowded environments, exploring various navigation frameworks based on deep reinforcement learning and their benefits over traditional simultaneous localization and mapping-based frameworks. Subsequently, we comprehensively compare and analyze the relationships and differences among three types of navigation: autonomous-based navigation, navigation based on simultaneous localization and mapping, and planning-based navigation. Moreover, the crowded environment includes static, dynamic, and a combination of obstacles in different typical application scenarios. Finally, we offer insights into the evolution of navigation based on deep reinforcement learning, addressing the problems and providing potential solutions associated with this emerging field.

Action recognition has gained great attention in automatic video analysis, greatly reducing the cost of human resources for smart surveillance. Most methods, however, focus on the detection of only one action event for a single person in a well-segmented video, rather than the recognition of multiple actions performed by more than one person at the same time for an untrimmed video. In this paper, we propose a deep learning-based multiple-person action recognition system for use in various real-time smart surveillance applications. By capturing a video stream of the scene, the proposed system can detect and track multiple people appearing in the scene and subsequently recognize their actions. Thanks to high resolution of the video frames, we establish a zoom-in function to obtain more satisfactory action recognition results when people in the scene become too far from the camera. To further improve the accuracy, recognition results from inflated 3D ConvNet (I3D) with multiple sliding windows are processed by a nonmaximum suppression (NMS) approach to obtain a more robust decision. Experimental results show that the proposed method can perform multiple-person action recognition in real time suitable for applications such as long-term care environments.

A method to locate the position of the target vehicle with the aid of infrared signal-direction discrimination for short-range vehicle-to-vehicle communications is proposed in this paper. With the use of two one-dimensional (1-D) signal-direction discriminators mounted on a vehicle with a lateral separation between them to measure the coming directions of the signal emitted from another target vehicle, the position of the target vehicle relative to this detecting vehicle can be located by triangulation. The aforementioned 1-D signal-direction discriminator is composed of two planar receiving modules forming a specific geometric structure, which leads to different responsivities of these individual receiving modules for a signal incident from a definite direction. The coming direction of the signal is determined by comparing the signal strengths received by these individual receiving modules. With a proposed hardware and software implementation, the position of the target vehicle can be accurately determined within a longitudinal range of 50 m. Owing to the restriction of the width of the vehicle, the separation of the signal-direction discriminators mounted on the vehicle cannot exceed 1.8 m. This leads to the consequence that in the remote region any small inaccuracy of the measured signal directions may cause very large position error. However, our measured results reveal a specific characteristic of the distribution of the measured positions for the target vehicle at a definite location. This allows us to design deliberate statistical algorithms to accurately extract the position of the target vehicle in the remote region. Using the simplest statistical algorithm, we are able to locate the position of the target vehicle with acceptable accuracy within a longitudinal range of 80 m. In this paper, we show the feasibility and applicability of this method. Furthermore, improvement is foreseeable with more deliberate design of hardware and software.

In this paper, we propose a method for improving the robustness of real-time facial expression recognition. Although there are many ways to improve the accuracy of facial expression recognition, a revamp of the training framework and image preprocessing allow better results in applications. One existing problem is that when the camera is capturing images in high speed, changes in image characteristics may occur at certain moments due to the influence of light and other factors. Such changes can result in incorrect recognition of the human facial expression. To solve this problem for smooth system operation and maintenance of recognition speed, we take changes in image characteristics at high speed capturing into account. The proposed method does not use the immediate output for reference, but refers to the previous image for averaging to facilitate recognition. In this way, we are able to reduce interference by the characteristics of the images. The experimental results show that after adopting this method, overall robustness and accuracy of facial expression recognition have been greatly improved compared to those obtained by only the convolution neural network (CNN).

A contest-oriented project for undergraduate students to learn implementation skills and theories related to intelligent mobile robots is presented in this paper. The project, related to Micromouse, Robotrace (Robotrace is the title of Taiwanese and Japanese robot races), and line-maze contests was developed by the embedded control system research group of the Department of Electronic Engineering, Lunghwa University of Science and Technology, Taiwan. It targets both those students who have to earn credits for a one-year special topics course and those who are just interested in making robots, and it is designed to motivate them to learn digital motion control, path planning, attitude correction, curvature detection and maze-solving algorithms. The students begin by getting acquainted with the development environment of microcontrollers, the characteristics of different sensors, and servomotor control techniques. Having learned these basic skills, they acquire further specific advanced skills and proceed to design their own mobile robots to compete in contests. The special topics course students' robots must pass examination by five teachers. Blogs and a wiki Web site for recording students' progress and experiences enhance the project's learning outcomes. Although not every student wins a prize in the contests, student feedback still shows that the contest-oriented project did motivate them to acquire the skills necessary to build and operate intelligent mobile robots.