Valter

ROS Software Maintainer: Alexey Maistrenko valtertherobot@gmail.com

The Valter is a mobile wheeled-based manipulation platform with anthropomorphic torso.

Valter was conceived as experimental platform for provisioning of various types of household services. As a matter of fact it is also a research and educational platform.

I see Valter mainly as an educational resource / platform for students.

Valter's autonomous behavior currently based on the following ROS packages:

rtabmap_ros, rgbd_slam, moveit!, move_base, dwa_local_planner, ORK find_object_2d, object_recognition, openni_kinect, openni_launch and more...

Contents

Valter

Specification

1. Components graph

Valter components graph

2. Features list

Feature	Value
Height	1.85 m
Weight	~ 110 kg
Case material	Duralumin
On-board computing means	RPi3 array, AT921SAM7S based controllers
Communication	Wi-Fi IEEE_802.11, Bluetooth 4.1, IR, Ethernet (for debug)
Power	12V car battry, 5V
Vision system	2 Kinect, stereo head camera system, navigation camera, back-view camera, manipulator camera, charger camera
Odometry motion sensors	Wheel encoders, compass, Kinect
Obstacle avoidance back-up system	IR Sharp proximity sensors array (GP2Y0A41SK0F), ultrasonic range finders array, IR bumper array (used also for IR control)
Microphones	1 high-gain mic for speech recognition, array of microphones for sound direction detection
Charger	Motorized / adjustable charger module for wall outlets
Sensing	over 30 force-measurement sensors
Arm lift up payload	6 kg (joints are in max torque applied position)
Lower manipulator paylod	1 kg (fully expanded)

Basic Configuration

Solidworks project

Kinematic specification

Joint	DOF \ Description
Platform	Unconstrained holonomic 2-wheeled
Torso	Yaw, Tilt
L&R Shoulder	Yaw
Upper Arm L&R	Pitch
Elbow L&R	Pitch, Yaw
Forearm L&R	Roll
Wrist L&R	Yaw, Pitch
Palm (gripper) L&R	6 fingers on each palm, 12 force sensors on each palm
Lower Manipulator (Third arm)	4 DOF, parallel gripper, 2 force sensors

Installation

Currently Valters' on-board sensor electronics and electromechanics are managed by custom Qt5 app which exposes TCP-IP gateways for intercommunication with ROS. This is exactly what is planned to translate entirely into native ROS interfaces (controllers).