S URV 2 AGV: a surveillance and survivability test-bed

Một phần của tài liệu A survivability framework for autonomous systems (Trang 244 - 248)

The SURV2AGV (AutonomousGroundVehicle forSurv-eillance andSurv-ivability) (Fig- ure C.17) is a small AGV designed for operation in both indoor cluttered environments and outdoor uneven terrain, and capable of carrying different sensing payloads. It is developed as part of the realization of theSurvivability Framework, and used as a test- bed for the experimental validation of the framework. This robot is designed around a reinforced aluminium chassis with damped suspensions and substantial articulation that makes it capable of outdoor, rough terrains.

Omnidirectional Reflector CCD Camera Bluetooth GPS Unit

IR Proximity Sensors 4-Wheel Drive/Steering

With Optical Encoders Processing Unit -Mini-ITX VIA CPU -802.11b/g WirelessLAN

Omnidirectional Reflector CCD Camera Bluetooth GPS Unit

IR Proximity Sensors 4-Wheel Drive/Steering

With Optical Encoders Processing Unit -Mini-ITX VIA CPU -802.11b/g WirelessLAN

Figure C.17: The Surv2AGV experimental framework (Autonomously GuidedVehicle for Surveillance andSurvivability), an all-terrain test-bed for surveillance and theSurvivability Framework.

C.3.1 Concept of operations

The SURV2AGV is designed to achieve the following operational objectives:

• To follow predetermined trajectories and provide a video feed back to the surveil- lance operations centre.

• To detect entities (e.g. human beings) intruding into a clearance radius.

• To observe and follow intruding entities until they leave the clearance radius.

• To return to base when power is depleted.

C.3.2 Hardware components

Small AGVs are typically limited in autonomy, given that many sensors required for autonomous functions are typically large, heavy and require significant amounts of power (Bruch et al., 2005). This limits the type of sensors that can be placed on a small AGV such as the iRobot PackBot, Koala Robot, and a man portable system such as the SPAWAR (SSC San Diego) URBOT (Bruch et al., 2005). In the SURV2AGV, many commercially-off-the-shelf (COTS) components and sensors are used (Table C.5) to optimize development time. In addition, many other components are developed from scratch if doing so is more efficient. This include the design and development of wheel odometers, and a sensor fusion board to combine proximity information from all the infrared sensors. The hardware architecture of the system is shown in Figure C.21, which depicts the different components and their inter-connections.

Table C.5: The SURV2AGV: Technical specifications

Component Specifications

Platform

Processor Mainboard VIA Mini-ITX, 1.2GHz C3 processor

Memory 1 GB DDR SDRAM

Chassis Tamiya TXT-1 monster truck kit (assembled) Dimensions length: 510mm, width: 385mm, height†: 297mm Wheels base: 335mm, tread: 280mm, diameter: 165mm

Tires Cross-country v-tread

Mobility 4-wheel driven (4WD), 4-wheel Ackermann steered (4WS) I/O Controller Acroname Brainstem GP 1.0

Motor Controller M’Troniks Sonik4 Electronic Speed Controller (ESC) Communications Netgear 802.11b/g wireless LAN

USB Bluetooth adaptor Sensors

Odometry 4 wheel encoders constructed from optical sensors Proximity 16 Sharp IR Proximity Sensors (8 LR + 8 SR)

Vision Omnidirectional camera system (developed in-house)

Heading Devantech Digital Compass

Positioning Holux GPSlim236 Bluetooth GPS receiver Software components

Operating System Windows XP Professional Service Pack 2 Cygwin v1.5.19 (0.150/4/2) 2006-01-20

†Chassis only, with no payload added. Height with payload varies.

Organization and physical layout

The system is divided into two distinct layers, as shown in Figure C.18. The topmost layer houses the computing hardware, namely the processor mainboard and its power supply unit, memory and storage media, wireless LAN and Bluetooth adaptors. The bottom layer comprises the I/O and motor controllers, sensor fusion board, Infrared proximity sensors, remote fail-safe unit (to cut off power to motors remotely in times of emergency), batteries, motors, and steering servos.

(a) Computing Layer (b) Motor Layer

Figure C.18: Hardware configuration for SURV2AGV.

Wheel Encoders and Odometer

The wheel encoders are constructed from scratch using a longitudinal encoder pattern mounted cylindrically around the wheel hub, and a photo reflector (Hamamatsu Model P5587) in a detection circuit. The configuration is shown in Figure C.19. Together they form a wheel odometer with a resolution of 70 counts per revolution. With a wheel diameter of 165mm, this translates to a linear resolution of 7.4mm.

(a) Side view (b) Front view

Figure C.19: Wheel odometer and encoder.

Infrared Proximity Sensors and Sensor Fusion Board

A total of 16 analog infrared proximity sensors (Sharp GP2D12 and GP2Y0A02YK) are placed around the vehicles (Figure C.20(a)). However, to read all analog voltages from these sensors require a total of 16 analog-to-digital channels. To save on such overhead, a sensor fusion board is designed and built to perform a weighted voltage summation of the 16 voltages into 4 channels, which can be fulfilled by the Brainstem controller.

The weights of summation are set with different resistor values. The final voltages are normalized in hardware (Figure C.20(b)).

(a) Infrared Sensors (front) (b) Sensor fusion board Figure C.20: Infrared sensors and sensor fusion board.

...

RF Failsafe E-STOP

Electric Motors

RS232

ADC

Servo I/O

I2C Bus

PWM

Digital I/O

USB

Firewire USB

PCMCIA

RF Failsafe Transmitter IR1

IR2

IR16 IR15 IR3

IR14 IR4

IR13 IR5

IR12

Infrared Sensor Interface

Electronic Speed Controller Brainstem GP

Controller

Servo, ADC, Digital I/O Digital Compass

Wheel Odometer Front Left Wheel Odometer Front Right Wheel Odometer Rear Left Wheel Odometer Rear Right

Bluetooth USB Adaptor

802.11b/g WirelessLAN Omnidirectional

Camera System

Trinocular Stereovision System

Processing System

Mini-ITX Via C3 1.2 GHz

GPS Receiver

Figure C.21: Hardware architecture for SURV2AGV.

Một phần của tài liệu A survivability framework for autonomous systems (Trang 244 - 248)

Tải bản đầy đủ (PDF)

(290 trang)