Mars Rover communications and autonomy Dr Anthony J H Simons (from NASA materials)

Mars Rovercommunications and autonomy

Dr Anthony J H Simons (from NASA materials)

2

Rover and Lander Different Configurations

Solo rover, or working with lander and/or orbiter? Landing: arrest via rocket, parachute, or airbag? Swarms: copters, biomorphs, penetrators

Different Control Strategies Signals direct to rover, or via orbiter, lander? Base, orbiter, lander, rover distribution of control Sensors: pressure, altimeter, laser, radar, vision Control: experiment deployment, data uplink

3

Example Missions NASA Pathfinder Sojourner

July-September 1997 Lander base station plus small rover http://www.jpl.nasa.gov/news/fact_sheets/mpf.pdf

NASA Mars Exploration Rovers January-April 2004 (but still active) Two independent rovers Spirit and Opportunity

landed at different locations http://www.jpl.nasa.gov/news/fact_sheets/

mars03rovers.pdf

4

NASA Pathfinder Sojourner

Image © Neil English, Exploring Mars, Pole Star Publications Ltd.

Antenna

Solar cells

Multi-wheel drive Steerable

front pair

5

Pathfinder Sojourner Lander

Airbags to cushion landing

Exit track for rover

Lander bounces (like ball)

Airbags deflate and shell opens

Lander has uplink

Image © Neil English, Exploring Mars, Pole Star Publications Ltd.

6

Slide © NASA, 2004. See http://robotics.nasa.gov/

7

Slide © NASA, 2004. See http://robotics.nasa.gov/

8

Slide © NASA, 2004. See http://robotics.nasa.gov/

9

NASA Spirit Rover (MER)

Image © Neil English, Exploring Mars, Pole Star Publications Ltd.

Stereo imaging

and navigation cameras

Direct-to-Earth uplink

Poseable instrument package

Multi-wheel drive

Special hazard

cameras

Strut for long reach

10

Slide © NASA, 2004. See http://robotics.nasa.gov/

11

Slide © NASA, 2004. See http://robotics.nasa.gov/

12

Deep Space Communications Distance to Mars

Closest to Earth: 54.5 million km Furthest from Earth: 401.3 million km

Signal times Based on c = 299,793 km/s ~ 3.03 minutes (Earth/Mars closest) ~ 22.31 minutes (Earth/Mars farthest)

Consequences Base cannot react in real-time Rover must act autonomously

13

Mission Management Base station (Earth)

Mission goals, priorities, master control Master data uplink, processing science results

Local station (Orbiter, Lander) Local area planning, local priorities, alternate tasks Global hazards, sandstorm warnings, rover safety local data uplink, local processing, data reduction

Rover Navigation, terrain following, obstacle avoidance Experiment selection, control, completion

14

Hardware Issues Satellite uplink

Need for Earth/Mars, Rover/Orbiter communications What hardware, comms. protocols, power rating?

Microcontrollers Small processors to read sensors and drive devices What memory, buses/ports, power rating, software?

Communications bus How many sensors, devices, moving parts to control?

Devices and sensors What devices/sensors? What registers to read/write?

15

Navigation Global Positioning System (GPS)

Could the Rover use this to find out its location? How many Orbiters/registration signals? How often/accurately measured? How important?

Tilt Sensors (Accelerometers) Compute velocity, position from known starting point

using internal acceleration sensors Integrate acceleration over time for velocity, velocity

over time for distance – but how to correct drift errors? Ultrasonic sensors

Echo location system for computing distance from target

Use in Martian atmosphere for obstacle avoidance?

16

Ultrasonic Sensors SRF08 ultrasonic sensor On-chip microcontroller

PIC determines distance from objects

Detects objects from 3cm – 6m

I2C bus communicates with external TINI

TINI concentrates on high-level control

Product image © Total Robots. SRF08 sensors available from Total Robots

17

Instrument Packages Navigation

Stereo navcams, hazcams, laser striper, ultrasound, inertial compass (no magnetic field)

Science 360 panoramic camera, HD cameras Spectrometers: infrared/thermal emission (carbon,

minerals), Moessbauer (iron-bearing properties) Rock abrasion tool Microscope (spores, bacteria) Wet science chemistry (lifesign reactions)

18

Software Issues Multi-tier

AI for high-level autonomous decisions Stereo vision algorithms for navigation Sensing, analysis and data compression

Reliability Triple-redundant voting system? Cosmic ray damage: reboot and/or reconfigure? Failsafe shutdown options

Communications Coordinate rover, lander, orbiter?

19

Slide © NASA, 2004. See http://robotics.nasa.gov/

20

Slide © NASA, 2004. See http://robotics.nasa.gov/

21

Slide © NASA, 2004. See http://robotics.nasa.gov/

22

Slide © NASA, 2004. See http://robotics.nasa.gov/

23

Slide © NASA, 2004. See http://robotics.nasa.gov/

24

Slide © NASA, 2004. See http://robotics.nasa.gov/

25

Slide © NASA, 2004. See http://robotics.nasa.gov/

26

Slide © NASA, 2004. See http://robotics.nasa.gov/

27

Slide © NASA, 2004. See http://robotics.nasa.gov/

28

Slide © NASA, 2004. See http://robotics.nasa.gov/

29

Future Missions NASA Phoenix Scout

Launched in 2007, Polar Lander Wet chemistry, water-finding, life?

NASA Mars Science Laboratory Launch in 2009, 10*payload of MER Skycrane rocket lander, nuclear power

Projected Biomorph Swarms Aerobot/rotorcraft, biomorph/micro-rovers and

subsurface penetrators Work as cooperating swarm, resilient to failures

30

NASA’s Phoenix Scout

Lander only

mission

Mission to northern

polar region

Subsurface water

ice?Wet water chemistry experimen

ts

Image © Neil English, Exploring Mars, Pole Star Publications Ltd.

31

NASA Mars Science Laboratory

Direct-to-Earth uplink

Much larger

rover (*10)

Image © Neil English, Exploring Mars, Pole Star Publications Ltd.

Nuclear powered

32

Stanford: Mesicopter Swarm

Swarm rotorcraft

Robust and

redundant

Cooperating agents

Image © NASA/DoD Second Biomorphic Explorers Workshop, JPL 2000.

Work by Ilan Kroo, Peter Kunz, Dept. of Aeronautics and Astronomy, Stanford University

33

Swarm Exploration

Image © NASA/DoD Second Biomorphic Explorers Workshop, JPL 2000.

Work by Ilan Kroo, Peter Kunz, Dept. of Aeronautics and Astronomy, Stanford University

34

ANTS Mission

35

Swarm Control Massively parallel system

How to predict all possible interactions? Cannot hope to test all behaviours System must be correct by design

Tools for understanding, specifying swarms Individual-based modelling (FLAME tool) models cellular automata Formal method: X-Machines specifies cellular automata

AnyQuestions?