ast-robotics · 2009-10-13 · 1 22 may 2009 ast-robotics public – not confidential ast-robotics...

1

22 May 2009

AST-Robotics

May 2009AST-RoboticsPublic – Not Confidential

UAVOutline1. Robotic Vision Systems

2. Inertial Platform

3. Sound Based Control

4. Robotic Communication

5. Motor Control

6. Unmanned Aerial Vehicle

7. Piezo actuators

Robotic Communication Robotic Vision Motor ControlMEMS Demo

Sound Based Control

2


Robotic Vision Systems Applications

Intelligent ToysVisual SLAM / 3D vision

Object recognition

Aerial imaging

Video surveillance and

intercom

Robotic aided surgery


Main FeaturesVery small Robotic Vision System (RVS)

30 x 30 mm System on Module (SoM)

this version can be directly soldered on your custom board

34 x 34 mm Socket version

ST-STR912FA Microcontroller (Arm966 @ 96MHz)

Free Development Toolchain

ST-VS6724 Camera Module 2Mpx

BMP or JPEG acquisition

16MB external burst PSRAM (100MHz)

Multiple Camera Synchronization

4 x 16 pins/pads for system connection

Application board

120 x 90 mm

Ethernet, USB, CAN, RS232, IR, MicroSD, I2C, SPI, GPIOs

RVS v. 1.0 RVS Module

Application Board

3


RVS v. 1.0

4 x 16 pins

for system

connection

ST-STR912FA

ARM966 @ 96MHz

ST-VS6724

2Mpx camera

16MB PSRAM

BURST


30 x 30 mm SoM version

SoM version30 x 30 mm

Surface Mount Device

4 x 16 pads

Socket version34 x 34 mm

Socket plug-in I/F

4 x 16 pins


RVS Development Board

USB

Switching Power Source

5-9V

LED HLJTAG

Ethernet 10/100

RS232

RS232

CAN

RVS

Module

uSD

LEGO

RS485 / I2C

Interface

IR

4


STLCam (ST LEGO Camera)

Host for Vision Module based on RVS 1.0

Communication:

USB

LEGO – I2C / RS485

Rear expansion connectors:

MII (Ethernet pins)

JTAG

GPIOs

Features:

Dual power USB / LEGO

Reset

An optimized RVS Main Board for LEGO Mindstorm NXT

4.8 cm

4.8 cm


Embedded Processing

ETH-RT streaming: 800x600 @ 30fps

RT processing up to 240x180 @ 30fps

Blob analysis (Contour points, Area,

Perimeter, Circularity, Outer box)

Face detection

Harris features extraction

Multiple camera synchronization

Face detection

Blob analysis

14874166Available instructions per pixel

640000032000003200000Available instructions per image

432004320019200Image size (pixel)

0.31%0.62%0.41%Acquisition overhead

153030Frame rate

240x180240x180160x120Image resolution

RVS performance table

RVS Capabilities

5


RVS PC Evaluation SW

Blob and graphical

settings Color blobs

Features

Multiple camera

management with a single

PC through Ethernet

Pre-testing of microcontroller

algorithms

Blob analysis

Face detection

Save and Load image files

Windows and Linux

supported

PC imaging environment to test the functionalities of an RVS module connected via Ethernet


Embedded blob analysis

Color rulesVisualization options Blob geometry

Embedded blob analysis algorithm

ported both on the RVS module and on the PC

evaluation software

6


Embedded face detection algorithm

can run either on the RVS module or on the

PC evaluation software

Embedded face detection


Embedded Harris

Harris Corner Extraction Matching algorithm

Harris Movement Vectors

Harris corner extraction algorithm can run either on the

RVS module or on the PC evaluation

software

7


Robots with Embedded Vision

AGV-JokerTableJoker

G2 ballFollower

LegoJokerRoombaBot


This system can feed synchronized images to a PC for low cost 3D avatar

reconstruction

Mobot Avatar

8






5. Motor Control


7. Piezo actuators

Robotic Communication Robotic Vision Motor Control

Sound Based Control

MEMS Demo


Inertial Platform: RIS v. 1.0Main Features

Very small Robotic Inertial System (RIS)

30 x 30 mm System on Module (SoM)

this version can be directly soldered on your custom board


ST-STM32F103RE Microcontroller

Arm CortexM3, 90DMIPS, 64KB RAM, 512KB Flash

Free Development Toolchain

ST-LIS331DLH

3 Axis, 16 bit digital accelerometer

ST-LY530AL

1 axes, 10 bit digital gyroscope

4 x 16 pins for system connection

Application board

48 x 48 mm

USB, RS-485, I2C, SPI, GPIOs

RIS Module

Application Board

9


RIS v. 1.0

STM32Arm CortexM3

90DMIPS

3 axis16 bit

accelerometer

Expansionconnector

1 axes10 bit

gyroscope

RIS Module (SoM version)

4 x 16 pins

for system connection

- 30 x 30 mm SoM version

- 34 x 34 mm Socket version Improved Localization

Flying stabilization

Robot stabilization

Lab Map

6 mt

Sta

rt

End

20 mt

Low-cost Map Building


RIS development board

Front side Back side

USB

GPIOs

JTAGResetRIS SoM

RS-485LEGO I/F

LED

48 mm

48 m

m

10


Lego-way Stabilization

Challenge:

Two wheels self-balancing robot

“The LEGO Segway”

Hardware:

Lego Mindstorm NXT

LY530AL - single axis yaw rate

MEMS gyroscope

LIS3LV02DL - 3 axis MEMS

accelerometer

STM32 - sensors fusion


Lego-way Architecture

StabilityControl

ENC

Gyro

Acc

STM32 NXT

1 varEKF

Pitchestimator

DriftFilter

11


G2 ball follower

Pb 12V battery

RVS Vision System

RIS Inertial Platform

Power Board Dual

DC-Motor

Centerstand


MEMS Remote Control

FWD - BWD LEFT - RIGHT

Bluetooth

Pitch

Roll

12






5. Motor Control


7. Piezo actuators


Sound Based Control

MEMS Demo


Audio Projects

Projects

Remote Control

Goertzel

Sound Localization

Triangulation

The System

3 microphones

STM32

Mobile robot

13


Sound Remote Control

Power

Freq

FWD - BWD LEFT - RIGHT


Mic 1

Sound Localization

α

Sound Source

Mic 2

Mic 3

β

AB

αα>0

α<0

14






5. Motor Control


7. Piezo actuators


Sound Based Control

MEMS Demo


Robotic Communication

MotivationIntelligence node 2

Intelligent vision

Intelligentrover

Intelligentmanipulator

Intelligence node 1

SM Slaves

Sniffer

Complex robotic tasks

easily solved by using

smart, robust and reusable robotic

modules

Real Time

Ethernet & SPI

ST is industry sponsor of

International Conference on Robot Communication and Coordination

15


Project Goal

Development of an Open Protocol for Real-Time communication of robotic subsystems

Readily available to the developer’s community

Portable over standard hardware (e.g. Ethernet)

To reduce costs

To simplify development

Deterministic behavior in any load condition of the network

Multi-master and capable of supporting a large number of nodes

Lightweight, in order to allow implementation on simple and low-cost nodes.

Scalable bandwidth, depending on underlying physical network


Access layer

Memory layer

Ethernet MAC(or any other comm. bus)

MTAP

User application Memory Layer

Manage data exchange

via remote memory

access

Access Layer

Manage bus contention

via token passing

Multi-master Tokened Access Protocol

Software stack

16


MTAP Protocol Testing

Test system in our lab

1 Super Master

(Lemote embedded PC based on ST

Loongson MIPS64@1GHz)

6 Slave nodes

(Raisonance microcontroller Dev board

based on ST STR9 ARM966E@96MHz)

1 Sniffer

(standard laptop PC)

1 Repeating hub

1 Ethernet switch

Test results


Exchanged data between

the master and each slave

2 bytes write

2 bytes read

6844FRTP=1KHz [N. slaves]

7885TSp-M [µs]

760494FRTP=100Hz [N. slaves]

1320TS-S [µs]

4653TM-S1 [µs]

rep. hubswitchTest type

MTAP Protocol: Test results over Ethernet

760 DC MotCtrl

Modules => 6080

controlled motors

at 100Hz

17


MTAP Protocol: Test results over SPI

Master1

Slave 1 Slave 2

SPI_SCKSPI_DATASPI_INT

47K 47KVcc

Slave n

Run on an STM32

SPI @ 18Mbps

CRC16 in HW

Cycle freq. up to 8KHz

New version: 2 wires

8KHzMax Cycle Freq.

40FRTP=1KHz [N. slaves]

65TSp-M [µs]

450FRTP=100Hz [N. slaves]

40TS-S [µs]

60TM-S1 [µs]

SPITest type


Promotes the convergence of robotics and communications, acting as a common forum for the two communities.

ST is Industry Chair

Sponsored Best Student Paper Award in 1st edition

2nd International Conference on Robot Communication and Coordination

Odense, Denmark, March 31 - April 2 2009

18






5. Motor Control


7. Piezo actuators


Sound Based Control

MEMS Demo


Features

ST STR9 Microcontroller (Arm9 @ 96MHz)

1 Brushless (BLDC) Motor connector (6 PWM)

8 DC Motors connectors, each with

1 PWM output

2 digital outputs for H-Bridge direction bits

1 analog input

1 quadrature encoder input (500 KHz sampled)

2 digital inputs for joint limit switches

8 RC servos outputs

8 channel configuration dipswitch

Speed and position control @ 1KHz

2 Ethernet 10/100 ports

Several communication peripherals (CAN, USB, RS232, I2C, SPI)

Real Time Ethernet Protocol Implementation (MTAP)

Free Development Toolchain based on Eclipse, GCC and GDB

Motor Control Module MC-ERT-188

19


MC-ERT-188

8 DC motor ports

8 CH dipswitch

8 RC servomotor OUTUSBSwitching

Power Source6-30V

2 Ethernet 10/100 ports

BLDC port

RS232

RS232 - CAN

JTAG

STR912FA

Eth hub/switch


AGV-Joker Demo(Autonomous Ground Vehicle)

Open Robotic Platform

Robotic arm

7 RC Servo Motors

G4 Rover

4 DC Motors

Eth

Motor control board

STR9 (96MHz) 8 DC motors - encoder 8 RC servos

D-Link DWL-G730AP Small lightweight

Wifi Bridge

Master Node

Laptop (2 GHz) Image processing Motor control Task control

WiFi

Eth

STR9 + 2Mpx camera Object tracking

Intelligent Vision

20


AGV-Joker Demo(Autonomous Ground Vehicle)

Open Robotic Platform

Robotic arm

7 RC Servo Motors

G4 Rover

4 DC Motors

Eth

Motor control board

STR9 (96MHz) 8 DC motors - encoder 8 RC servos

D-Link DWL-G730AP Small lightweight

Wifi Bridge

Master Node

Laptop (2 GHz) Image processing Motor control Task control

WiFi

Eth

STR9 + 2Mpx camera Object tracking

Intelligent Vision






5. Motor Control


7. Piezo actuators


Sound Based Control

MEMS Demo

21


UAV - Applications

Surveillance and Imaging Dangerous Environments

Swarm Robotics (cooperation) Indoor Navigation & Mapping


UAV - Basic movements

4 basic movements:

Throttle (U1)

Roll (U2)

Pitch (U3)

Yaw (U4)

22


UAV - Dynamic system

6 DOF rigid body dynamics

Rotational kinematics

Basic movements derivation


UAV - Control law implementation

==

=

+−=

ZZ

YY

XX

IU

IU

IU

mUgZ

/

/

/

/)cos()cos(

4

3

2

1

ψϑφ

ϑφ

&&

&&

&&

&&

−−=

−−=

−−=

−−=

ψψψψϑϑϑϑ

φφφφ

ψψ

ϑϑ

φφ

&&&

&&&

&&&

&&&

Dd

P

Dd

P

Dd

P

DZd

PZ

KK

KK

KK

ZKZZKZ

)(

)(

)(

)(

( )( )

( )( )( )

−−=

−−=

−−=

−−+=

ψψψϑϑϑ

φφφϑφ

ψψ

ϑϑ

φφ

&

&

&

&

Dd

PZZ

Dd

PYY

Dd

PXX

DZd

PZ

KKIU

KKIU

KKIU

ZKZZKgmU

)(

)(

)(

)cos()cos(

)(

4

3

2

1

−−−

−=

ΩΩΩΩ −

4

3

2

1

1

24

23

22

21

00

00

U

U

U

U

dddd

lblb

lblb

bbbb

Simplified dynamic system Output movement signals

Basic control rule

Output propellers’ speeds

23


UAV - Simulator

Simulation and 3D visualization through Matlab/Simulink

Altitude and attitude references

Dynamic Simulator

3D Visualization


UAV - System connections

UAV

Data download

Real Time ControlUAV block diagram

R/C receiver IR sensor

Sonar sensor

IMU sensor

Power moduleBattery

24


UAV – Simulation and Real Flight






5. Motor Control


7. Piezo actuators


Sound Based Control

MEMS Demo

25


Rotational Piezo Motor


Our Robotic Fleet

G2 Ball Follower

RoombaBot

F4

Lego-Way

LegoJokerAGV-Joker

Table-Joker

ast-robotics · 2009-10-13 · 1 22 may 2009 ast-robotics public – not confidential ast-robotics...

Documents