2.2.1 Mars Rovers

This segment of the program supports the development of robotics to satisfy the planned requirements for exploration of the surface of Mars. These plans call for robotic reconnaissance and exploration systems traversing the Mars terrain. During such missions robots will explore potential landing sites and areas of scientific interest, place science instruments, and gather samples for analysis and possible return to Earth. The robotic systems required for these operations will require high levels of local autonomy, including the ability to perform local navigation, identify areas of potential scientific interest, regulate on-board resources, and schedule activities, all with limited ground command intervention. The objectives of the tasks within this segment of the program are to develop these abilities, as well as conduct research into mobility systems, miniature mechanisms, planning, and on-board navigation. Specific applications are to the Mars Pathfinder, Mars Surveyor Network project and other programs planned by the Space Science user community.

Rover Technology Program

Robotic Systems Controlled by Multiple Agents

Lander Manipulation

Remote Geologist

Controlled Descent

Technology Transfer Roadmap Details

Rover Technology Program

Introduction/Rationale: The rover technology program is designed to develop a technology base from which will emerge a capability to enable extensive robotic science exploration of select areas of Mars. This technology base will greatly expand the current Mars Pathfinder and Surveyor microrover performance in the areas of goal identification, increased vehicle mobility, intelligent terrain navigation, obstacle/hazard avoidance, fault management, more flexible science data gathering and manipulation of science instruments, and rover/lander functionality partitioning.

Focus and Direction:

This technology development program combines both research and actual system demonstrations as a means of pushing the state of existing autonomous vehicle technologies while maintaining flight program relevance. While the primary focus is research, the planned system demonstrations provide a means of testing the robustness of the technology components within a viable mission scenario/environment. The following major milestones are planned:

FY '94 Autonomously traverse 100m of rough terrain within sight of the lander.

FY '95 Perform survey of a 20mx20m area, executing 10 science sorties (i.e., rover based site imaging/soil sampling) with only 5 command uplinks. Elimination of 5 of the major failure modes identified in FY94 will be demonstrated/documented; survey will include images of site areas occluded from the lander (non-line-of-sight navigation), as part of each sortie the rover will perform at least one goal identification or sample acquisition at designated sites with return to the lander vicinity. Rover survey will be conducted in variable terrain (terrain ranging from .6-2 large obstacles per sq. m.).

FY '96 9-10 Kg vehicle, thermal insulation, multiple science target confirmation, multiple sampling and instrument pointing/deployment modes, traverse with return to landers

FY '97 Miniaturized vehicle (goal is 5 Kg); non line of sight traverse, component capabilities tested in thermal vacuum chambers

Points of Contact:
Wayne Zimmerman
(818)354-0234
wayne.zimmerman@jpl.nasa.gov

Brian Wilcox
(818) 354-4625
brian.wilcox@jpl.nasa.gov

Robotic Systems Controlled by Multiple Agents

The purpose of the Behavior Control for Multiple Agencies task is to develop reliable reactive control techniques for indoor and outdoor navigation and obstacle avoidance, to develop high-level navigation skills (including position disambiguation, map making, map following and resource handling), and to develop techniques for efficiently integrated reactive and high-level techniques.

Reactive robots typically have excellent performance, but have difficulty taking instructions or relaying information, and therefore can deal with only limited task complexity without extensive programming. Strategic robot planners solve the communications problem, but are very inefficient. This task will enable the merger of the two technologies, taking the best features of each, blending together strategic planning with real-time performance of behavior controlled reactive systems. This will be done by letting the reactive system control the robot, and letting the strategic system reprogram the reactive system. If successful, the results of this task will allow real-time robots to use and make maps, follow instructions, and give users more insight into the robot's control and actions.

In FY 1995, this task has been incorporated into the Rover Technology task, while work on the Boadicea research vehicle continues.

Point of Contact:
Rod Brooks
(617) 253-5223
brooks@ai.mit.edu

Lander Manipulation

This task develops a fully tested and characterized advanced technology manipulator arm and core drill for deployment from a small planetary lander. The advanced manipulator will be tested in a simulated Mars environment to insure suitability for future mission applications. The goal of this task is to apply advanced actuator and materials technology to increase manipulator dexterity while decreasing power, volume and mass. In addition, this activity is coordinated to integrate advanced control technologies developed by the MET/Sample Evaluation/Acquisition Testbed (SEAT) task. Baselined against existing flight qualified technologies applied to a Mars Global Surveyor (MGS) sized manipulator arm (see parameters described for MGS in FY '95 deliverable below), a 30% volume reduction, a 60% mass reduction and 30% power reduction have been established as end goals. Future lander dexterous manipulator applications include: sample acquisition, core drilling, science instrument surface placement and recovery. Technology developed by this task will resolve serious low-mass, low-power manipulator issues facing scientists and designers of future small lander based missions. In this regard, potential users will be actively involved in defining the baseline functional parameters and mission applications.

Focus and Direction:

FY '95 An MGS sized manipulator with three degree of freedom will be designed, fabricated, and demonstrated. Current MGS performance parameters include; Reach: 1.5 m, Mass: less than 5 kg, Operating Power: less than 10 watts, Stowed Volume: 15 cm X 15 cm X 40 cm. The MGS manipulator will be assembled using qualified but unproven Mars Pathfinder MicroRover actuators and tested to establish a performance benchmark for future development. Compared to existing flight qualified technologies, a 10-20% reduction in mass and power is anticipated. During FY '95, a collaborative effort will be established with industry and universities to develop advanced actuators and a core drill. In addition, manipulator requirements will be established to provide flexible yet positive performance parameters to obtain end-of-program goals.

FY '96 By the end of the second year, the baseline arm will be upgraded and demonstrated using an advanced SEAT control system. Actuator technology developed during this period is expected to provide an additional mass and power reduction of 10-15% over the previous baseline. Actuator testbed components will be designed and tested under Mars operating conditions (-100 C, 10 torr, dust). During this period, a new optimized structure, control system and coredrill will be designed to allow fabrication, integration and test to occur the following year. In addition to lightweight actuators, advanced composites, bearings and lubrication will be evaluated for specific applications.

FY '97 At the close of this year, a newly designed and improved five degrees of freedom arm will be demonstrated with advanced actuators and materials technologies. Concepts for interchangeable tools and zero force power connectors will be developed. During this period, the force and feedback technologies developed by the SEAT task will be integrated with the manipulator and tested. Low temperature operation and feedback loops will be verified. In addition, a coredrill prototype will be fabricated and separately characterized prior to manipulator integration.

FY '98 As the fourth year ends, the advanced manipulator, coredrill and feedback control system will be fully integrated. The manipulator, and coredrill will be optimized and demonstrated in a Mars simulated environment. Compared to the three degrees of freedom baseline, a 50 % volume reduction, a 50% mass reduction and a 30% power reduction is anticipated. Primary activities of this period will center around systems optimization, characterization and environmental testing. Initial vibration characterization of the coredrill and feedback to the control system are expected be of significant interest. Additional tools such as scoops, rock picks or instrument pick-and-place adaptors will be prototyped and tested to explore the performance envelope of the advanced manipulator, system. At the end of this period, a summary report will be completed to insure that the developed technology is accessible to support multi-task operations for future mission use.

Point of Contact:
Curtis E. Tucker
818-354-6133
curtis.e.tucker@jpl.nasa.gov

Remote Geologist

This task will prototype a mobile robotic science platform capable of remote scientific exploration with minimal human supervision. Time delay in planetary communication and the limitation of one uplink/ downlink cycle per day does not permit scientists to guide remote sampling and analysis in real-time. Furthermore, constraints on communication bandwidth may not allow scientists to receive and reduce the raw data and decide a new course of action in a reasonable amount of time. Technical focus of this task will therefore be the development of an "intelligent science system", an autonomous control system which will:

	Execute high-level commands in terms of science activities such as core and analyze a specified rock
	Automatically reduce data to classify samples and produce "executive summaries" of findings allowing scientists to quickly access results and direct continued operation
	Simultaneously carry out exploratory type commands, for example looking for a specific type of material in samples
	Conduct multiple science tasks and recover from failure enabling independent operation for prolonged periods of time

This technology will help maximize science data gathered by enabling the remote system to direct current activities while still allowing scientists overall supervision of the mission. Planetary scientists will be an integral part in the task identifying required science activities as well as data reduction and classification methods. A mobile platform with prototype science instruments and sampling devices will be developed enabling testing of the "remote scientist" under the control of planetary scientists.

Focus and Direction:

FY '95 Interaction with planetary scientists to develop requirements and design of intelligent science system. Pre-prototype low mass/power (2kg/10w) sub-surface sampling device for Champollion Surface Science Package (Rosetta Comet Encounter Mission) and demonstrate sub-surface sampling to 50cm in proto-cometary material.

FY '96 Pre-prototype sample handling mechanism for Champollion sub-surface sampling device developed in FY '95 and integrate on testbed with one science instrument such as visible/IR spectrometer. Demonstrate single command capability for sample acquisition, handling and analysis.

FY '97 Integration of second science instrument such as Mossbauer spectrometer to aid in material characterization and development of automatic sample classification capability. Outdoor test of science package on mobile platform demonstrating sample acquisition, in-situ analysis, and automatic classification of multiple samples.

FY '98 Develop automatic correlation of data between multiple instruments to improve ability to identify materials and implement data summary methods. Outdoor demonstration of mobile science platform sampling, analyzing and returning data summaries under remote control of scientists at universities/NASA centers.

FY '99 Development of mission level control enabling multiple science activities to be carried out for autonomous operation. Two week demonstration mission in challenging remote environment with 1 uplink/downlink per day to scientists supervising mission from JPL (no human interaction at remote site).

Points of Contact:
Richard Welch
(818)354-7042
richard.v.welch@jpl.nasa.gov

Controlled Descent

This task will develop and demonstrate technologies in low-cost robotic vision for landing including:

Simple, low-cost image processing

Range rate (time to impact) determination

Horizontal translation determination

Optical based position & attitude determination

Hazard detection and avoidance algorithms

In cooperation with the MET* program, this task will lead to a single low mass/power autonomous landing system that can in real-time: determine its position and movement relative to the surface of a planetary object; track and avoid detrimental hazards; and land safely in rugged terrain. State-of-the-art advances will include:

	miniaturization and mass reduction made possible by combining several sensor functions (attitude determination, Horizontal velocities, time to impact, local-navigation, hazard detection) into a single real-time autonomous system
	addition of new types of sensors and data processing to enhance knowledge of the vehicle position and velocity during descent
	adding lateral control and hazard avoidance goes beyond capabilities demonstrated by Viking and other planetary landers operating in actual missions
	achieving a capability that goes beyond Viking with a much smaller and less costly vehicle
	achieving Apollo-like control in lateral maneuvering and descent, without human involvement

Focus and Direction:

FY '95 The main technical accomplishment in FY '95 is a real time optical based computation of time to impact, and to establish the corresponding experimental proof-of-concept testbed for demonstration of realistic descent image processing. The MET program will prototype a small testbed vehicle, and passive vertical descent will be demonstrated in multiple experimental trials. This task will provide this vehicle with technology in low-cost robotic vision to support the landing experiments. These experiments verify that the image processing, range rate (time to impact), feature tracking, and real-time-computing can be accomplished during descent in a realistic environment.

FY '96 The FY '96 system will greatly improve the capability of the FY '95 experiments by developing optical-based range and attitude determination algorithms for local terrain maneuvering.

FY '97 On-board vision system determines the lateral velocity of the lander with respect to the local terrain, and issues commands to the lander maneuvering system to reduce the lateral velocity to a small value.

FY '98 Data from the on-board sensors is used to detect hazards on the planet surface, and on-board algorithms maneuver the vehicle to avoid these hazards.

FY '99 Low-cost vision system is used to detect hazards from real-time imaging. The MET landing maneuvering system issues commands to the lander to avoid hazards. Capability to land in rugged terrains is demonstrated.

* This task is complementary to the Mars Exploration Technology Program efforts in microlander vehicle development by providing a low-cost vision system for hazard avoidance and precision landing.

Point of Contact:
George Powell
818-354-4067
george.e.powell@jpl.nasa.gov