Kui Chen

Fundamental development of a virtual reality simulator for four-arm disaster rescue robot OCTOPUS

This paper presents a virtual reality (VR) simulator for four-arm disaster response robot OCTOPUS, which has high capable of both mobility and workability. OCTOPUS has 26 degrees of freedom (DOF) and is currently teleoperated by two operators, so it is quite difficult to operate OCTOPUS. Thus, we developed a VR simulator for training operation, developing operator support system and control strategy. Compared with actual robot and environment, VR simulator can reproduce them at low cost and high efficiency. The VR simulator consists of VR environment and human-machine interface such as operation-input and video- and sound-output, based on robot operation system (ROS) and Gazebo. To enhance work performance, we implement indicators and data collection functions. Four tasks such as rough terrain passing, high-step climbing, obstacle stepping over, and object transport were conducted to evaluate OCTOPUS itself and our VR simulator. The results indicate that operators could complete all the tasks but the success rate differed in tasks. Smooth and stable operations increased the work performance, but sudden change and oscillation of operation degraded it. Cooperating multi-joint adequately is quite important to execute task more efficiently.

Development of a prototype electrically-driven four-arm four-flipper disaster response robot OCTOPUS

In a previous study, we have developed a four-arm four-crawler disaster response robot called ‘OCTOPUS’. Advanced disaster response robots are expected to be capable of both mobility, such as entering narrow spaces over unstructured ground, and workability, such as preforming complex debris-removal work. We have confirmed experimentally that the four arms could make the robot perform complex tasks while ensuring stabilization when climbing steps while the four flippers could make it traverse rough terrain while avoiding toppling over when conducting manipulation task. OCTOPUS, renamed as H-OCTOPUS, is oil-hydraulically driven to perform outdoor demolition of heavy debris, and is teleoperated by two operators. In this study, we develop a prototype electrically-driven OCTOPUS, called E-OCTOPUS, to manipulate various light-objects mainly indoors such as valve operations in nuclear power plants. For reducing the size and weight while maximizing task performance, we introduced a mutual complementary strategy between its arms and flippers. To validate the capability of E-OCTOPUS, we performed preliminary experiments involving climbing high steps and manipulating and cutting wires by cooperating the four arms and four flippers. The results indicated that E-OCTOPUS could complete the tasks by coordinating its four arms and four flippers.

Analysis of operation strategy in a multi-operator control system for four-arm disaster response robot OCTOPUS

Disaster response robot with four arms and flippers OCTOPUS has high mobility and task-execution capabilities. Owing to the higher number of degrees of freedom, OCTOPUS is controlled by two operators, however this kind of robots is inherently difficult to be operated. To design easy-to-use human machine interfaces and intelligent control systems, we need to analyze and quantify a reasonable operation strategy in multi-operator control systems. Thus, three different types of essential disaster response tasks were conducted by using OCTOPUS and we analyzed results of operations and work performance, by focusing on each operator and each pair. As the results, we derived basic operation strategies as follow; operators with higher number of simultaneously-operated joints (Ns) can control OCTOPUS more smoothly, and pairs with higher rate of cooperated operations (Rc) can finish tasks more efficiently. We also found that Ns and Rc can be used to quantify operational skills. Revealed strategies and parameters could be useful to design new human-machine interface and intelligent control system.

A preliminary study on a groping framework without external sensors to recognize near-environmental situation for risk-tolerance disaster response robots

This paper proposes a basic near-environmental recognition framework based on groping for risk-tolerance disaster response robot (DRR). In extreme disaster sites, including high radiation and heavy smog, external sensors such as cameras and laser range finders do not work properly, and such sensors may be broken in accidents in the tasks. It is hoped that DRRs can continue to perform tasks, even if the external sensors cannot work, and at least, they can safely evacuate from the site. In this preliminary study, for recognizing near environments without using external sensors, we proposed a groping method. In this method, a robot actively touches the environment using arms or other movable parts, records the contact information, and then reconstructs a three-dimensional local map around the robot by the detected information, e.g., robot arms position and reactive force. The proposed groping system can recognize the existence of three situations, such as an object, step, and pit, and those geometry, by exploring the designated space using arms. The groping strategy was designed considering both robot specification, time limitation, and required resolution. Experiments were performed using four-arm and four-crawler robot OCTOPUS. The results indicate that the proposed framework could recognize step, pit, and object, and calculate the position and size of the object, and confirm that the robot successfully removed the object on the basis of groped data.

A semi-autonomous compound motion pattern using multi-flipper and multi-arm for unstructured terrain traversal

Disaster response crawler robot OCTOPUS has four arms and four flippers for better adaptability to disaster environments. To further improve the robot mobility and terrain adaptability in unstructured terrain, we propose a new locomotion control method called compound motion pattern (CMP) for multi-limb robots like OCTOPUS. This hybrid locomotion by cooperating the arms and flippers would be effective to adapt to the unstructured terrain due to combining the advantages of crawling and walking. As a preliminary study on CMP, we proposed a fundamental and conceptual CMP while clarifying problems in constructing CMP, and developed a semi-autonomous control system for realizing the CMP. Electrically-driven OCTOPUS was used to verify the reliability and correctness of CMP. Results of experiments on climbing a step indicate that the proposed control system could obtain relatively accurate terrain information and the CMP enabled the robot to climb the step. We thus confirmed that the proposed CMP would be effective to increase terrain adaptability of robot in unstructured environment, and it would be a useful locomotion method for advanced disaster response robots.

Usability Test in Different Types of Control-Authority Allocations for Multi-Operator Single-Robot System OCTOPUS

Four-arm four-flipper disaster response robot called OCTOPUS, which has 26 degrees of freedom (DOF), has been developed to engage in complex disaster response work. OCTOPUS adopts a multi-operator single-robot (MOSR) control system, and two operators manually control the robot. The pattern of control-authority allocation (CAA) for two operators largely affects performance of MOSR systems, so this study conducts usability tests in various CAAs, and derives a reasonable CAA pattern. So far, there are no uniform standard allocation rules for flippers and crawlers in multi-limb robots like OCTOPUS, therefore we specify five CAA patterns for investigation. Three fundamental tasks that whole body of the robot must be cooperatively controlled were conducted to test the usability in each CAA pattern. From the results of analysis, we found that pattern 1, which one operator controls front two arms, flippers, and crawlers, and another operator controls remaining back parts, had best scores for all tasks, as well as pattern 2 had the best distribution of workload between two operators.