Vision-based techniques for intelligent vehicles in heterogeneous road environments are gaining significant attention from researchers and industrialists. Unfortunately, the mechanisms in this domain suffer from limited performance due to scene complexity, varying road structure, and improper illumination conditions. These challenging situations may lead an intelligent vehicle into dangerous situations such as collisions or road accidents and may cause higher mortality. The application of intelligent methods and other machine learning techniques for road surface classification is little explored in the existing literature. Thus, we propose a convolutional neural network-based road classification network (RCNet) for the accurate classification of road surfaces. This procedure includes the classification of five major categories of road surfaces: curvy, dry, ice, rough, and wet roads. The experimental results reveal the behavior of the proposed RCNet under various optimizer techniques. The standard performance evaluation measures have been used to test and validate the proposed method on the Oxford RobotCar dataset. RCNet achieves classification accuracy, precision, and sensitivity of 99.90%, and 99.97% of specificity. Results of implemented work are significantly higher than available state-of-the-art techniques and show accurate and effective performance in the complex road environment.

Deep reinforcement learning has the advantage of being able to encode fairly complex behaviors by collecting and learning empirical information. In the current study, we have proposed a framework for reinforcement learning in decentralized collision avoidance where each agent independently makes its decision without communication with others. In an environment exposed to various kinds of dynamic obstacles with irregular movements, mobile robot agents could learn how to avoid obstacles and reach a target point efficiently. Moreover, a path planner was integrated with the reinforcement learning-based obstacle avoidance to solve the problem of not finding a path in a specific situation, thereby imposing path efficiency. The robots were trained about the policy of obstacle avoidance in environments where dynamic characteristics were considered with soft actor critic algorithm. The trained policy was implemented in the robot operating system (ROS), tested in virtual and real environments for the differential drive wheel robot to prove the effectiveness of the proposed method. Videos are available at https://youtu.be/xxzoh1XbAl0.

Research on robots that can be used for communication with humans has become popular in recent years. Communication robots should ideally be as small as an infant in order to reduce the user’s feeling of threat. In addition, non-verbal communication (such as gestures) is also important in facilitating smooth interactions between humans and robots. There are currently a few communication robots that are small sized and can generate hand gestures. In this paper, we propose a small robot hand, which is optimized for gesture communication by using a shape memory alloy (SMA). The SMA employed is a Ti–Ni alloy, which is used as an actuator. The SMA shrinks when it transforms into the austenite phase at temperatures higher than the transformation temperature. When it is in the martensitic phase at a lower temperature, it is expanded by an external force. Each finger of the robot hand is driven by an individual SMA wire. The specifications of the small robot including the hand size, operation angles in each finger joint, response times and power consumption were determined according to the human finger and existing small communication robots. These required specifications have been fulfilled by carefully designing the geometry and heating/cooling power control. A questionnaire-based survey was also conducted with a robot hand. The five-finger hand was successfully shown to generate recognizable symbolic gestures.

The Industry 4.0 paradigm promises shorter development times, increased ergonomy, higher flexibility, and resource efficiency in manufacturing environments. Collaborative robots are an important tangible technology for implementing such a paradigm. A major bottleneck to effectively deploy collaborative robots to manufacturing industries is developing task planning algorithms that enable them to recognize and naturally adapt to varying and even unpredictable human actions while simultaneously ensuring an overall efficiency in terms of production cycle time. In this context, an architecture encompassing task representation, task planning, sensing, and robot control has been designed, developed and evaluated in a real industrial environment. A pick-and-place palletization task, which requires the collaboration between humans and robots, is investigated. The architecture uses AND/OR graphs for representing and reasoning upon human-robot collaboration models online. Furthermore, objective measures of the overall computational performance and subjective measures of naturalness in human-robot collaboration have been evaluated by performing experiments with production-line operators. The results of this user study demonstrate how human-robot collaboration models like the one we propose can leverage the flexibility and the comfort of operators in the workplace. In this regard, an extensive comparison study among recent models has been carried out.

Autonomous navigation in complex environment is an important requirement for the design of a robot. Active SLAM (simultaneous localization and mapping) combining, which combine path planning with SLAM, is proposed to improve the ability of autonomous navigation in complex environment. In this paper, fully convolutional residual networks are used to recognize the obstacles to get depth image. The avoidance obstacle path is planned by Dueling DQN algorithm in the robot’s navigation; at the same time, the 2D map of the environment is built based on FastSLAM. The experiments show that the proposed algorithm can successfully identify and avoid moving and static obstacles with different quantities in the environment, and realize the autonomous navigation of the robot in a complex environment.

Magnetic actuation techniques and microrobots have attracted considerable interest due to their potential applications in biomedicine. Interventional techniques have emerged as a minimally invasive approach to treat a wide range of vascular diseases. The current practice of interventional procedures is, however, limited by manual control of interventional devices and time-consuming procedures. Moreover, fluoroscopy is considered as an essential part of the procedure today despite posing many limitations for patients and physicians. Recently, various microrobotic solutions have been proposed for vascular interventions, including advances in magnetic navigation systems and magnetically steerable catheters and guidewires, which have shown potential benefits such as reduced radiation doses, improved access to difficult-to-reach and tortuous anatomy. This paper reviews the commercial magnetic actuation systems and magnetically actuated interventional microrobots that have been developed by academic research groups and medical companies worldwide, outlining their capability, applicability as well as limitations. We further address the challenges and future prospects of the research toward clinical acceptance of magnetic interventional technologies.

In this research, we studied the mating tolerance of various plug-in cable connectors and provide a mating tolerance dataset of 70 different connectors. This dataset will be highly advantageous to industries for wire harness assembly tasks using robots. Understanding the mating tolerance is crucial for automating the mating process because it is closely related to the control specifications of a robotic manipulator. Our system uses a 2-finger Robotiq adaptive gripper attached to a 6 degree-of-freedom industrial robot (ABB Robotics) to test the mating process of wire harness assembly tasks. In addition, we use 70 types of wire harness connectors with different numbers of pins widths, lengths, and thicknesses, and various shapes, to test the mating tolerance. The results indicate that the connector mating tolerance of our dataset is more generous than the repeatability of conventional industrial manipulators, and further demonstrate the suitability of the position control methods to wire harness assembly tasks.

Previous learning-based methods that rely on 2D laser data to classify indoor robot locations into semantic classes were successful in distinguishing between rooms and corridors. However, the classification accuracy remained low for doorway locations. We propose a semantic place classification method that uses a rule-based doorway detection algorithm followed by a classification scheme that models training data through either K-means clustering or learning vector quantization. We conducted extensive experiments on the Freiburg 79 dataset and compared our method to previous semantic place classification algorithms. The doorway detection algorithm we propose significantly increases the classification accuracy for doorway locations as compared to the state-of-the-art performance. We applied our method, trained on the Freiburg 79 dataset, to Freiburg 52 and ESOGU datasets in order to demonstrate its generalization ability.

Loop closure detection (LCD) is crucial for the simultaneous localization and mapping system of an autonomous robot. Image features from a convolution neural network (CNN) have been widely used for LCD in recent years. Instead of directly using the feature vectors to compute the image similarity, we propose a novel and easy-to-implement method that manages features from a CNN via a novel approach to improve the performance. In this method, the elements of feature maps from the higher layer of the CNN are clustered to generate CNN words (CNNW). To encode spatial information of CNNW, we create word pairs (CNNWP) that are based on single words to improve the performance. In addition, traditional tricks that are used in methods that are based on bag of words (BoW) are integrated into our approach. We also demonstrate that the feature maps from lower layers can be used as descriptors to conduct local region matching between images. Via this approach, we can perform geometric verification for possible loop closures, similar to BoW methods, in our approach. The experimental results demonstrate that our method substantially outperforms state-of-the-art methods that directly use CNN features for LCD.