In this paper we describe a continuous development and evaluation of multi purpose platform robot AERO via robot competitions. We participated to 4 robot competitions requiring different task with human size semi humanoid robot AERO, has same basic structure. For many types of competition tasks, the robot was designed as simple, robust, small and lightweight, furthermore it can be change parts for another situation of task. The versatility of the robot platform was increased from continuation of task analyisis, consideration of requirements, and implementation. In this paper we describe a continuous integration of robot design, cycle of task analysis, implemetation and integration, via robot competition.

The use of field robots can greatly decrease the amount of time, effort, and associated risk compared to if human workers were to carryout certain tasks such as disaster response. However, transportability and reliability remain two main issues for most current robot systems. To address the issue of transportability, we have developed a lightweight modularizable platform named AeroArm. To address the issue of reliability, we utilize a multimodal sensing approach, combining the use of multiple sensors and sensor types, and the use of different detection algorithms, as well as active continuous closed-loop feedback to accurately estimate the state of the robot with respect to the environment. We used Challenge 2 of the 2017 Mohammed Bin Zayed International Robotics Competition as an example outdoor manipulation task, demonstrating the capabilities of our robot system and approach in achieving reliable performance in the fields, and ranked fifth place internationally in the competition.

Robots are designed to tackle specific problems or are designed as a multipurpose platform that achieves a concept. In this letter, we approach the hardware design problem from practice and evaluation through numerous competitions including the Darpa Robotics Challenge, the Amazon Picking Challenge, and other competitions that were held in Japan. The paper aims to answer the question what hardware design would we achieve if we continuously refine designs through competitions? We show that competitions lead to a more compact but capable platform design that is durable and has more end effector solutions. We also point out the limitations and benefits of using competitions for designing robot platforms. Our platform began from a human-inspired design for teleoperation and has evolved to mix both human-like and industrial robot structure.

The sequential online tuning for controller gains is required for the continuous action of the riding into parallel two-wheeled scooter and the speed governing after riding by humanoid robot. The implemented controllers are different between the riding and the speed governing, and these tuning strategies are also different. In particular, the riding requires the immediate tuning in the short riding phase and the speed governing requires the accurate tuning to regulate the speed of humanoid robot. To the above requirements, this paper proposes the Sequential Online Learning Control (SOLC) method composed of the cascade connection of SGD-based open-loop Learning Control (SLC) and Mini-batch-based closed-loop Learning Control (MLC). SLC contributes the damping gain online tuning for the foot torque control during execution of riding, and MLC contributes the PID gains online tuning for the speed governing control. Finally, we show the validity of SOLC through the sequential experiment of riding and speed governing for parallel two-wheeled scooter by life-sized humanoid robot HRP2-JSK.

This paper proposes the online learning controller for PID gain tuning to regulate the speed of robot on parallel two-wheeled electric scooter. We define this speed regulation as “speed governing behavior”. The proposed system based on Iterative Feedback Tuning (IFT) contributes the online continuity during learning control not to interrupt and reset/restart the controller. PID gain update during control has the problem that the integral gain change affects the divergence of control input. To this problem, we propose the digital integrator which modifies the previous integrated value of control error. This proposed integrator solves the impulsive control input without any anti-windup or reference shaping. In conclusion, this paper shows the experimental results based on the online learning control and demonstrates the speed governing behavior by life-sized humanoid robot. We finally aim the driving of parallel two-wheeled electric scooter by speed control of humanoid robot.

In order to enable humanoids to pick a unstable object, it is necessary to track the position and pose of the object in real time by the image and point cloud from camera.But the conventional methods to track objects by using 3D Optical flow are able to track only the position of objects moving by translational motion.In this paper, we propose the real time object position and pose tracking method using segmentation of point cloud in addtion to using 3D optical flow. We applied this method to control the picking motion of humanoid robot and realize picking unstable objects.

We present a field robot platform HRP2G with humanoid upper body and a high power mobile wheeled base that focus on outdoor tasks. We assemble the hardware of both parts and create a software bridge between the humanoid upper body and the mobile wheeled base through popular robotics software ROS. For the humanoid upper body, we use HRP2 robot and for the mobile wheeled base we buy individual parts and create our own control unit with both hardware and software. The kinematics of the robot is well modelled and we design the corresponding control law. We will apply this robot platform to participate in the second task of MBZIRC(Mohamed Bin Zayed International Robotics Challenge) in 2017. In the experiment we show how our platform accomplished certain tasks with customize designed gripper and sensor equipments, which demonstrates the feasibility of our integrated system.