Robot-RTT

Project card

• Project Name: Study and implementation of a self-localization system for mobile robot navigation on rough terrain.
• Authors: Juan Camilo Gómez Cadavid (kmilo17pet@gmail.com) and Nelson David Muñoz Ceballos (ndmunoz@elpoli.edu.co)

• Academic Year: 2011-2012

• Degree: Master

• Jde Version: jde-4.3

• SVN Repository: [1]

• Tags: Rough terrain, AUV, IMU, Inertial localization

• Technology: C, C++, JdeRobot, GTK+, dsPIC-Microcontroller based custom hardware, IMU-MEMs sensors

• State: Developing

• Source License:

The mobile robot platform

Mobile robot description

The RTT mobile robot

The mobile robot RTT (view Fig 1) is a rover-type research prototype Rocker-Bogie suspension [1][2], six-wheel independent, grouped in two sets of three wheels on each side of the vehicle and secured by means of an articulated structure. The rear wheels are linked to the robot's body by a rigid arm known as Rocker, and the same is fixed by means of a pivot, a second arm known as the bogie, which holds the middle and front wheels. A differential mechanism that connects the two rocker arm to the body of the vehicle, keeps the RTT balanced even when the trains are at different heights. Overall, the differential system further the articulated structure, seek to ensure that all six wheels are always kept in contact with the floor, allowing a permanent traction.

Figure 1: Rover-type research prototype mobile robot (RTT)

Problem

The mobile robots in specific applications as in agriculture, surveillance, search and functions on industrial environments must face to the difficult conditions of navigation that present the terrain where they operate. Although suspension mechanisms allow the operation of mobile robots in rough terrain, the problem is that the navigation systems start to generate unwanted behaviors because their localization techniques are affected by changes in the rugged terrain and constant noise in the sensorial measurements. The objective of this project is to study and implement an auto-localization system for mobile robot navigation in rough terrain. The system to implement aims to tackle the difficult navigation conditions that occur on uneven terrain where they can make few assumptions when it is desirable to plan the navigation of the robot. We expect that the system will provide the robot the ability to self-locate within their environment and the ability to navigate along the terrain answering to the spatial alterations that should be arise.

Project Objectives

• Propose a local self-localization system for the navigation of the mobile robot navigation RTT on rough terrain, taking into account the constraints of its mechanical structure.

For this main objetive, is necessary:

• Analyze different techniques of local self-localization applicable to robot RTT.
• Define the hardware and software architecture to accomplish the tasks of self-localization on rough terrain.
• Design and implement necessary hardware and software to estimate the robot location in the environment, taking into account the structural constraints of the robot RTT.
• Validate the operation of self-localization system by implementing autonomous navigation tasks on rough terrain.

Development

Hardware and Software architecture

• Sensorial System

Proprioceptive Sensors: The RTT has sensors to inform several aspects of the state in which the robot is, these are: -Current sensors (indirect torque measure)

-Orientation sensors

-Battery level sensors

-Encoders - Wheels Angular speed

-Attitude Sensors

Exteroceptive sensor: Supported by a SRF02 ultrasonic sensor array [3], located taking into account the reflective characteristics typical of an ultrasound wave and considering the maximum detection focus that includes the angular width and the minimum and maximum distance detected. Similarly, it is considered the effects of crosstalk noise referred to the crossing of signals between ultrasonic sensors. The configuration used in the RTT is composed of 7 ultrasonic sensors located as shown in Fig 2. Each sensor uses an I2C communication system. The distance measurements are managed through a USB interface, by a dedicated computer with a data acquisition system custom designed, using a 16-bit embedded microcontroller.

Figure 2: Distribution of ultrasound sensors. (Circular panorama and detection focus).

Inertial sensor: Three CruizCore® R1001E fully self-contained MEMS digital gyroscopes for measuring heading angles were used. Each uses a digital UART-USB board as communication interface with the dedicated computer, and are located in each axes of the robot’s body frame. The R1001E has 50Hz bandwidth and precisely measures angular rates up to ± 100 °/sec, it can also measure rates up to ± 150 °/sec with lesser accuracy. There is also a low cost Freescale® MMA7260® 3-axis accelerometer as second source of redundant information for tilt estimation and an electronic compass for alternative estimate of the robot pose. This sensors, allow estimate the robot attitude. The attitude is usually expressed in terms of three special angles known as “Euler angles”. The angles are φ, θ, and ψ, which are usually referred to as roll (sometimes also called “bank angle”), pitch (also called “elevation”), and yaw (also called “heading” or “azimuth”) respectively. Rates of rotation of the body frame relative to the navigation frame can be expressed in terms of the derivatives of the Euler angles, [4][5]. The inertial measurement unit (IMU) is shown in Fig 3.

Figure 3: RTT Inertial measurement unit (IMU)

• Hardware Control System

The control unit has an onboard-compact dedicated computer, with an Intel ® Atom ™ N270 1.60 GHz, with sufficient processing speed to perform autonomous navigation algorithms and control systems. Sensor data and control signals are handled by the computer through USB, using two data acquisition interfaces (DAQ).

The primary DAQ interface board (DAQ-1) is formed by a digital signal controller (DSC) dsPIC33FJ256MC710 [6], which is responsible for sending control signals to each of the traction motors through low-level PID controllers, providing stability and ability to function in a variety of terrain. In addition has the function of collecting and processing the signals from the proprioceptive sensors.

The second DAQ interface board (DAQ-2) uses a PIC18F4455 microcontroller [6], which aims to receive information from a set of exteroceptive sensors that provide data and inertial environment and state of the robot respectively.

Communication interfaces with the central processing system are established via the USB port, while other peripherals for sensor managing are performed using different communication buses such as I2C, SPI and RS232.

Additionally, it profits from computer's wireless card for making a call via WiFi, for monitoring, teleoperation and platform programming. The WiFi allows communication range up to 200 meters and the possibility of extending this distance using routers in repeater configuration [7].

Figure 4: RTT Hardware Architecture

References

[1] M. Thianwiboon, V. Sangveraphunsiri, R. Chancharoe. Rocker-Bogie Suspension Performance. Robotics and Automations Laboratory, Chulalongkorn University, Bangkok, Thailand 10330. 2001.

[2] Suojun Li; Haibo Gao; Zongquan Deng; , "Mobility performance evaluation of lunar rover and optimization of rocker-bogie suspension parameters," Systems and Control in Aerospace and Astronautics, 2008. ISSCAA 2008. 2nd International Symposium on , vol., no., pp.1-6, 10-12 Dec. 2008

[3] J. A. Echeverri A. S. Montoya C. Implementación de un módulo de sensórica para un robot de terrenos irregulares. Instrument & Control Engineering PFC. Politécnico Colombiano Jaime Isaza Cadavid. Medellin, 2009.

[4] Boney I.A et’al, Advantages of the modified Euler angles in the design and control of PKMs, Proceeding of the 3rd Chemnitz Parallel Kinematics Seminar/ 2002 Parallel Kinematic Machines International Conference, Chemnitz, Germany, April 23--25, pp. 171-188, 2002

[5] Ojeda, L.; Borenstein, J.; "FLEXnav: fuzzy logic expert rule-based position estimation for mobile robots on rugged terrain," Robotics and Automation, 2002. Proceedings. ICRA '02. IEEE International Conference on , vol.1, no., pp. 317- 322 vol.1, 2002

[6] Microchip Technology INC. dsPIC33FJ256MCXXX & PIC18F4455/4550/ Data Sheet, 2003. Available on: www.microchip.com

[7] TELEPIEZA.COM, Dispositivos Inalámbricos WIFI para una Red LAN (Punto de Acceso y Punto de Extensión), Disponible en: www.telepieza.com/wordpress/2008/05/14/dispositivos-inalambricos-wifi-para-una-red-lan-punto-de-acceso-y-punto-de-extension/