MTR: THE MULTI-TASKING ROVER

A New Concept in Rover Design

Antonios K. Bouloubasis, Gerard T. McKee, Paul M. Sharkey, Peter Tolson

Active Robotics Laboratory, School of Systems Engineering, University of Reading,

Whiteknights, Reading, RG6 6AY, UK

Keywords: Mobility, Internal/External Re-configurability, Modularity, Upgradeability.

Abstract: In this paper we present the novel concepts incorporated in a planetary surface exploration rover design that

is currently under development. The Multitasking Rover (MTR) aims to demonstrate functionality that will

cover many of the current and future needs such as rough-terrain mobility, modularity and upgradeability.

The rover system has enhanced mobility characteristics. It operates in conjunction with Science Packs (SPs)

and Tool Packs (TPs) – modules attached to the main frame of the rover, which are either special tools or

science instruments and alter the operation capabilities of the system.

1 INTRODUCTION

On July 4, 1997 a new era for space robotics and the

exploration of Mars began when the Pathfinder

mission successfully delivered the Sojourner rover

to the Red Planet. Following that, in 2005 the two

MER rovers, Spirit and Opportunity, traversed many

kilometres and took hundreds of pictures giving

much more information than their ancestor. All three

missions are the initial phase of a plan with the

ambition of the eventual human habitation of Mars.

So far the increasing numbers of missions to

Mars have provided important information about the

Martian climate and geology. Scientists are using

this information to locate areas of interest in the

surface of the planet and following that, robotic

rovers can be deployed to obtain ground truth

(NASA Mars Exploration Study Team, 1998). Areas

of interest are often very difficult to reach requiring

a rover to have increased rough terrain mobility

capabilities (Schenker, et al, 2000). The return of

samples to Earth for further examination with

equipment which is too sensitive to be sent to space

is also necessary (Garvin, 2003),

(Huntsberger, et al,

1999). After a number of sites have been evaluated,

the best in terms of recourses and topographic

location will be selected for the construction of

habitats to support human presence on the Red

Planet. Mobile rovers will be used throughout these

missions.

In order to carry out the tasks mentioned above,

different robotic mechanisms and rover designs need

to be employed. In this paper we propose a novel

new concept for robotic rovers, namely the Multi-

Tasking Rover (MTR). The idea behind the design

of the MTR is the fusion of all these systems into

one. This is accomplished with the construction of a

main rover system (Fig. 1), with enhanced mobility

and re-configuration capabilities, which will be the

carrier of different modules each dedicated to a

specific task. The rover will not only offer mobility

to these modules, but in combination with a

particular module will acquire unique characteristics

transforming its role and functionality. The modules

can be either Science Packs (SPs) or Tool Packs

(TPs) - the current design supports the deployment

of two Packs on each MTR.

For example a Science Pack can be a particular

type of spectrometer and a Tool Pack a scoop

mechanism. A scenario could be the transportation

Figure 1: The Multi-Tasking Rover. (MTR).

176

Bouloubasis A., Mckee G., Sharkey P. and Tolson P. (2006).

MTR: THE MULTI-TASKING ROVER - A New Concept in Rover Design.

In Proceedings of the Third International Conference on Informatics in Control, Automation and Robotics, pages 176-181

DOI: 10.5220/0001218401760181

 SciTePress

of the packs to a selected site, the acquisition of

samples using the scoop and in situ testing by the

spectrometer. Now assume that samples are needed

from a particular depth under the Martian surface.

The MTR will re-configure by placing the

Spectrometer and the Scoop Packs to a storage

location, pick-up two TPs, a robotic mole and a

deployable solar panel, move them to the desired

location, deploy them, connect them such that the

solar panels provide power continuously to the

robotic mole and leave them to that location until the

samples are taken.

The MTR approach assumes that the packs

have in-built control systems and can operate once

deployed independently from the rover.

Communication links between the packs and the

MTR will be established when required. The

advantage of this approach is that instead of sending

a large number of different rovers to perform a

variety of tasks, a smaller number of MTRs could be

deployed with a large number of different SPs and

TPs, offering greater functionality at a reduced

payload.

The remainder of the paper is organized as

follows. Section II describes the electromechanical

design of the MTR system. Section III outlines the

rover electronic and sensory systems. Section IV

gives a description of the behaviours that will be

implemented and the architecture, under which they

will be integrated. Finally, section V provides a

summary and conclusions.

2 ELECTROMECHANICAL

SYSTEMS

A key element in the development of a modular, re-

configurable, multitasking system like the MTR is

the development of complex mechanisms that will

enable the principles of operation to be

demonstrated. The MTR requires a total of 14

motorized actuators. It comprises of the following

subsystems: drive/steering system, active suspension

and base unit. An SP and/or a TP will be constructed

as well so that fundamental principles of operation

are demonstrated. According to the nature of the

Pack, this may introduce further axes of control. A

stereo camera system will also be integrated in the

design at a later stage.

The four-wheeled rover will achieve a

maximum speed of 7cm/sec, which is delivered

through a motor/gearbox combination incorporated

within each wheel. The aluminium wheels measure

175mm in diameter and the rims are covered with a

rubber tire for maximum traction. Each of the

wheels is independently steered giving the rover the

highest mobility possible. The MTR can traverse

forward/backward, turn on the spot, take hard/soft

turns and crab to any direction maintaining the

orientation of the body. The rotation of each of the

wheels is restricted to ±185 degrees by limit

switches.

Figure 2: Demonstrating internal re-configurability of the

Carrier.

The Active Suspension Mechanism (ASM)

serves as means of not only providing rough terrain

stability, by re-allocating the vehicle’s centre of

mass (Figure 2), but also gives the basic mobility to

the main body for reaching, grasping and deploying

any of the Science/Tool Packs that need to be

employed for a given task. The ASM comprises of a

pair of shoulders and each of these in turn comprises

a pair of legs; at the end of each leg is a steerable

wheel. Each shoulder’s angle is adjustable between

0 to 188 degrees allowing the main body to move

up/down (300mm travel) and modify its roll angle

(±26 degrees). This is accomplished using a linear

actuator located within each leg. If so desired the

shoulder’s angle can be adjusted by altering the

configuration of only one of the legs.

Figure 3a, illustrates a 3-D model one of the

shoulders fully extended (lower position) and Figure

3b shows an assembly of all of the parts that have

been made to date. This configuration gives unique

motion characteristics to the rover’s body, enhancing

internal re-configurability. Each leg also houses a

Lithium-Polymer (Lipoly) battery and the associated

low-level controller. The four Lipoly packs situated

on the legs in conjunction with four more located

inside the chassis of the MTR give a total power

capacity of 22V at 9 Ah.

The topology of the MTR’s suspension is similar

to that of JPL’s SRR2K (Schenker, et al, 2000), but

configuration and functionally differ greatly. The

two shoulders are linked via an active differential

drive mechanism in order to obtain contact of all

four wheels with the ground. This is accomplished

with the Main Frame Rotation Mechanism (MFRM).

The main body resides between the two shoulders,

houses the differential mechanism, the on-board

high level controller and provides means of support

for the deployment of two Packs. The axis that links

MTR: THE MULTI-TASKING ROVER - A New Concept in Rover Design

177

the two shoulders through the differential is also the

axis of rotation of the body. The MFRM comprises

two actuators which provide the ability to adjust the

pitch angle (±720 degrees rotation) of the body

Figure 3: Three-dimensional model of the ‘shoulder’ (a),

and the parts that have been currently made (b).

of the MTR in order to maintain a constant

orientation to the horizontal when needed. The

MFRM mechanism also offers centre of mass re-

allocation for extra stability and body pitch angle

adjustment for the sake of operation of any Packs (a

TP might have to operate vertically or at an angle

e.g. a drill). Finally the MFRM gives the ability to

the main frame to pick-up a Pack no matter its

orientation; the roll angle can be controlled via the

suspension and the yaw angle can be determined

through the steering/drive system. Twelve actuators

in total control the subsystems mentioned above.

The mechanisms or instruments that can be

incorporated within a Pack are limited by the

maximum allowable size of the Pack and the

lifting/transportation capability of the MTR. The

maximum volume for a Pack is limited to 5litres and

its weight should not exceed 3kgs. Nonetheless this

configuration offers great external re-configurability

since an appreciable number of devices can be

deployed within the given constraints.

The MTR provides a set of mounting points on

its body to support two Packs. In order to simplify

the MTR design, each Pack encapsulates a locking

mechanism, necessary for stabilising it on the

rover’s body. Another advantage is the

upgradeability of the system since by sending new

Packs future needs of space exploration can be

satisfied. An absolute necessity is of course a

standard interface between the MTR and the

modules.

3 ELECTRONICS AND SENSING

The electronics system comprises two subsystems.

The first, the low-level controller, is built around the

Microchip PIC controller and a number of different

peripherals. It has the responsibility of motor PID-

servo control, as well as obtaining the sensors’

feedback to be utilised by local, low-level

behavioural loops, or the higher-level controller (the

second subsystem) when necessary.

SUSPENSION

DRIVE

Leg #4

Leg #2

Leg #3

Body Controller

High Level Controller

Leg #1

Controller

MFRM

Pack Alignment

Sensors

STEERING

RS485

Figure 4: Electronic & Electromechanical Subsystems on

the Carrier.

Modularity is a key design goal. The low-level

controller is divided into five smaller subsystems.

Each leg will comprise a small network of five PICs,

three motion controllers and two additional general

purpose controllers that will be used for functions

like A/D conversion, sonar reading and other low-

level functions that may be required. The 5

subsystem is located in the body of the MTR and

will be in charge of the actuators that govern the

operation of the MFRM (active differential and body

rotation). This controller will provide all the

necessary feedback for alignment of the MTR with

respect to a Pack.

The second subsystem, an on-board high-level

controller, will be connected with all the modules

ICINCO 2006 - ROBOTICS AND AUTOMATION

178

through an RS485 bus allowing a sufficiently large

number of devices to be part of the loop. The

platform that will be employed to perform the high-

level control functions is still under investigation.

The options range from the Nano-ITX and the

Soekris, to the very small Gumstix. The basic

topology of the high and low-level controllers,

together with the main electromechanical systems is

shown in Figure 4.

A Pack can have a controller of equivalent or

higher processing power, as the situation and

functionality demands. De-centralized control has

been the basis for fast response through parallel

processing and is not limited within the physical

boundaries of the MTR. If more processing power is

Figure 5: The Complete System.

required in order to carry out a given task, it can be

obtained from a TP/SP with enhanced processing

capabilities. A wireless Ethernet connection will

offer a fast data communication path between the

Control Station, the MTR and any of the Packs

(Figure 5).

A variety of different sensors must be employed

so as to obtain all the required feedback and assist

the function of the two subsystems. Two-channel

quadrature encoders will be employed for the PID

control of the rotating elements. Temperature

sensors will inform the controllers on the status of

the motor driver chips and current sensing will

provide the necessary force feedback. Strain gauges

incorporated within the steering system of each

wheel will monitor the contact forces with respect to

the ground and assist the operation of rough terrain

stability behaviours.

Sonar sensors, also based on the steering

brackets, as well as on the MTR’s body, will utilize

the pan rotation of the wheels and the tilt rotation of

the frame (MFRM) to support obstacle avoidance

behaviours. A two-axis inclinometer will provide

feedback on the roll and pitch of the vehicle. A pair

of GPS receivers will be employed (one on the MTR

and one on the Pack) in order to obtain rough

estimates for the position of the vehicle with respect

to the Pack. An RF receiver on the MTR will work

in combination with an RF transmitter on the Pack to

enable it to approach the Pack. Alignment and

grasping of the Pack will be performed using

infrared receiver/transmitter pairs, in conjunction

with digital compasses on both the MTR and the

Pack.

Note that many of the sensory devices mentioned

above cannot operate in a space environment.

Nonetheless alternatives exist that do. Usage of

cheaper systems allows the principles of operation of

the MTR system to be demonstrated.

4 BEHAVIOUR & CONTROL

SYSTEM ARCHITECTURE

Many functions of the system will be behaviour-

supported as this offers fast response times and

simplifies the overall control task. In avoiding

hazardous situations e.g. tipping over whilst

traversing on the sides of a crater in order to acquire

samples, or while cooperatively transporting an

extended payload (Bouloubasis, et al, 2005),

reflexive responses can be employed. Direct control

of the fourteen actuators would necessitate

enormous processing power and is not considered as

an option.

The MTR offers many opportunities for

behaviour based control. The Obstacle Avoidance

Behaviour (OAB) will utilise the ultrasonic sensors’

output to provide collision-free traversal when

enabled. In combination with an on-board digital

compass the maintenance of the course of traversal

will be ensured in case the vehicle must deviate from

its original path to avoid a collision.

As mentioned above, internal re-configurability

by means of re-allocating the rover’s centre of mass

through the suspension system (ASM) and the

frame’s rotation around the axis that links the two

suspension shoulders (MFRM) aims to offer rough

terrain stability. The Stability Enhancement

Behaviour (SEB) will obtain feedback from the

strain gauges located in the steering system and the

two-axis inclinometer to decide whether the

shoulder angle and/or the base pitch need to adjust to

accommodate differential altitude changes in the

rover’s local terrain.

In some cases it may be required e.g. for the

operation of a SP/TP, to maintain a particular

orientation of the body with respect to the

horizontal. The Orientation Maintenance Behaviour

(OMB), using the information obtained from the

inclinometer, will adjust the pitch angle of the body

to ensure that the desired orientation is maintained

during traversal of varying slope terrains.

Control Station

SP/TP

Wireless Ethernet

MTR: THE MULTI-TASKING ROVER - A New Concept in Rover Design

179

The approach to a Pack remote from the MTR

will be accomplished using an RF beacon. Once

enabled, the Pack Approach Behaviour (PAB) will

alter the rover’s velocity vector to point towards the

RF source. Following that the Pack Docking

Behaviour (PDB) will align the rover’s body with

the Pack. This will be established using a pair of

digital compasses in conjunction with infrared

transmitter/receiver pairs situated on the contact

faces of the MTR and the Pack. Once alignment is

verified, the ASM module will lower the body of the

MTR so as to obtain contact with the Pack. Once

contact is established the Pack will utilise the

mounting points offered by the MTR and physically

couple the two systems (Figures 6a-b).

The integration of a Pack may introduce

additional behaviours in the control system. For

example a Tool Pack may contain a manipulator

used for the cooperative transportation of extended

payloads (Bouloubasis, et al, 2003). It has been

demonstrated that a number of behaviours and

specialized sensory systems can be incorporated for

the completion of such a task (Bouloubasis, et al,

2005). Figures 6a-e shows a typical sequence of

actions that the MTR must perform in order to

acquire and utilize the Manipulator TP: dock to the

Pack, acquire contact and grasp it, lift it and rotate it

to the desired height and angle of operation, and

finally deploy it.

A single architecture must integrate all the

behaviours mentioned above and more importantly

any new behaviours introduced to serve the

operation of any SP/TPs. A multilayered architecture

(Brooks, 1986) assumes the addition of ‘levels of

competence’ to the existing ones, to achieve further

functionality. In an upgradeable, multi-functional

system like the MTR, this translates to either a very

complex hierarchical structure at the low level,

capable of accommodating additions in the higher

level of the architecture, or new levels that suppress

the lower ones. The designer cannot possibly predict

the behaviours that future Packs may require to

operate and therefore cannot predict ways that these

will interact with the existing MTR.

The single-layered Ego Behaviour Architecture

(EBA), (Lewis, et al 1997) is comprised of a number

of behaviours, which operate autonomously and

independently of each other. Each behaviour is

developed separately, tested and then integrated into

the existing architecture using an elementary

summation function (Fig 7). This facilitates the

design and suits the operation of the MTR since it

fulfils the need for uncomplicated assimilation of

new behaviours in the existing architecture. Another

advantage of the EBA is that the arbitration

mechanism allows cooperation or competition of

two or more behaviours, the emergent response

being the resultant effect rather than a single

behaviour. For example if PAB, OAB and SEB are

enabled, the rover will traverse towards the Pack

(RF source), change the velocity vector to avoid any

obstacles, maintain course, and at the same time

adjust its centre of mass through ASM and MFRM

systems to account for rough terrain.

Each Behaviour has an associated Ego (Figure 7).

The behaviour takes as input its current Ego Status

(Active or Inactive) together with any system

variables, e.g. sensory information, commands from

control station, etc. and produces a Desired

Response. A Desired Status signal, which depends

on a number of constrains, is also produced,

indicating whether the behaviour should be active or

inactive. The Ego of a behaviour compares the

system’s emergent response to that of the associated

behaviour and changes the Ego-controller gains in

order to gain control. When a behaviour fails to gain

control it becomes inactive (resigns).

d. Rotating the frame

e. The Manipulator TP is deployed

c. Lifting the Pack

a. Docking b. Securing the Pack

Figure 6: The MTR deploys the Manipulator Tool Pack.

ICINCO 2006 - ROBOTICS AND AUTOMATION

180

Behavior 1

Behavior 2

Behavior n

Ego 1

Ego 2

Ego 3

System

Inputs

Resultant

Response

Figure 7: The Ego-Behaviour Architecture.

The EBA is based on the concept that a number

of behaviours can work cooperatively and/or

competitively at the same time. In the example

mentioned, three behaviours are enabled to assist the

completion of a task, namely the safe approach to a

Pack. The PAB when enabled will direct the rover

towards the Pack. This works competitively with

OAB which will alter the vehicle’s direction when

an obstacle is detected. The sonar sensors’ input to

the behaviour will change the Desired Status to

Active. In this case the gain of the OAB Ego-

controller will be higher than that of the PAB and so

OAB will take over. When the obstacle has been

bypassed the behaviour will indicate a Desired

Status of Inactive; the Ego-controller of OAB will

resign giving the control back to PAB. In the same

example OAB and SEB will work cooperatively

towards safe traversal of the rover to the target area.

5 SUMMARY & CONCLUSIONS

The work presented in this paper outlines innovative

rover systems design concepts which could be

integrated to existing or future planetary surface

exploration rover designs. Emphasis has been given

in this paper to the overall design of the Multi-

Tasking Rover system. The MTR focuses mainly on

modularity and upgradeability, which are enhanced

by re-configurability (internal & external) of the

structure. Science Packs (SPs) and Tool Packs (TPs)

provide varying functionality to the MTR system.

Figure 8 shows the MTR equipped with two Packs: a

Manipulator TP that lifts small rocks and a

Spectrometer SP that examines soil underneath.

Our current work focuses on the design and

construction of mechanical and electronic systems

for the MTR. A Pack is also being designed in order

to demonstrate the fundamental principles of

operation of the MTR. The integration of the EBA

with the associated behaviours will follow the

completion of the electronics architecture. Further

developments will include incorporation of a stereo

camera system and vision-guided navigation.

Figure 8: A Manipulator TP is used in combination with a

Spectrometer SP to examine areas of interest under small

rocks.

REFERENCES

NASA, Human Exploration of Mars: The Reference

Mission (Version 3.0 with June, 1998 Addendum) of

the NASA Mars Exploration Study Team, Exploration

Office, Advanced Development Office, Lyndon B.

Johnson Space Center, Houston, TX 77058, June,

1998.

P. S. Schenker, et al, “Reconfigurable robots for all terrain

exploration”, in Proceedings of SPIE Vol 4196,

November 2000.

James B. Garvin, Daniel J. McCleese, NASA's Mars

Exploration Program: Scientific Strategy 1996 - 2020,

6th International Conference on Mars, April, 2003.

T. L. Huntsberger, et al, "Sensor-Fused Autonomous

Guidance of a Mobile Robot and Applications to Mars

Sample Return Operations", Part of SPIE Conference

on Sensor Fusion and Decentralized Control in

Robotic Systems II, Boston, Massachusetts,

September, 1999.

Antonios K. Bouloubasis and Gerard T. McKee,

Cooperative Transport of Extended Payloads,

Proceedings of ICAR 2005, Seattle, pp. 882-887,

2005.

A. K. Bouloubasis, G. T McKee, P. S. Schenker, A

Behavior-Based Manipulator for Multi-Robot

Transport Tasks, in the IEEE International Conference

on Robotics and Automation (ICRA) 2003, Taipei,

Taiwan, May 2003, pp. 2287-2292.

R. A. Brooks, "A Robust Layered Control System for a

Mobile Robot", IEEE Journal of Robotics and

Automation, 1986.

M. G. Lewis, P. M. Sharkey, “A plug and play architecture

for emergent behaviour in robot control”, Proceedings

Configuration an Control Aspects of Mechatronics,

Ilmeneau, Germany, September 1997.

MTR: THE MULTI-TASKING ROVER - A New Concept in Rover Design

181