Autonomous Navigation Performance During the Hartley 2 Comet Flyby
No abstract available
EPOXI mission↗
Engineering topics
Bhaskaran, Shyam
Publications and source records attributed to Bhaskaran, Shyam.
Autonomous Navigation Performance During The Hartley 2 Comet Flyby
On November 4, 2010, the EPOXI spacecraft performed a 700-km flyby of the comet Hartley 2 as follow-on to the successful 2005 Deep Impact prime mission. EPOXI, an extended mission for the Deep Impact Flyby spacecraft, returned a wealth of visual and infrared data from Hartley 2, marking the fifth time that high-resolution images of a cometary nucleus have been captured by a spacecraft. The highest resolution science return, captured at closest approach to the comet nucleus, was enabled by use of an onboard autonomous navigation system called AutoNav. AutoNav estimates the comet-relative spacecraft trajectory using optical measurements from the Medium Resolution Imager (MRI) and provides this relative position information to the Attitude Determination and Control System (ADCS) for maintaining instrument pointing on the comet. For the EPOXI mission, AutoNav was tasked to enable continuous tracking of a smaller, more active Hartley 2, as compared to Tempel 1, through the full encounter while traveling at a higher velocity. To meet the mission goal of capturing the comet in all MRI science images, position knowledge accuracies of +/- 3.5 km (3-?) cross track and +/- 0.3 seconds (3-?) time of flight were required. A flight-code-in-the-loop Monte Carlo simulation assessed AutoNav's statistical performance under the Hartley 2 flyby dynamics and determined optimal configuration. The AutoNav performance at Hartley 2 was successful, capturing the comet in all of the MRI images. The maximum residual between observed and predicted comet locations was 20 MRI pixels, primarily influenced by the center of brightness offset from the center of mass in the observations and attitude knowledge errors. This paper discusses the Monte Carlo-based analysis that led to the final AutoNav configuration and a comparison of the predicted performance with the flyby performance.
flyby↗
Autonomous Navigation for Deep Space Missions
Navigation (determining where the spacecraft is at any given time, controlling its path to achieve desired targets), performed using ground-in- the-loop techniques: (1) Data includes 2-way radiometric (Doppler, range), interferometric (Delta- Differential One-way Range), and optical (images of natural bodies taken by onboard camera) (2) Data received on the ground, processed to determine orbit, commands sent to execute maneuvers to control orbit. A self-contained, onboard, autonomous navigation system can: (1) Eliminate delays due to round-trip light time (2) Eliminate the human factors in ground-based processing (3) Reduce turnaround time from navigation update to minutes, down to seconds (4) React to late-breaking data. At JPL, we have developed the framework and computational elements of an autonomous navigation system, called AutoNav. It was originally developed as one of the technologies for the Deep Space 1 mission, launched in 1998; subsequently used on three other spacecraft, for four different missions. The primary use has been on comet missions to track comets during flybys, and impact one comet.
navigation↗
Stardust-next : Lessons Learned from a Comet Flyby Mission
The Stardust-NExT (New Exploration of Tempel) mission, a follow-on to the Stardust prime mission, successfully completed a flyby of comet Tempel-1 on 2/14/11. However there were many challenges along the way, most significantly low propellant margin and detection of the comet in imagery later than antici-pated. These challenges and their ramifications forced the project to respond with flexibility and ingenuity. As a result, the flyby at an altitude of 178 km was nearly flawless, accomplishing all its science objectives. Lessons learned on Stardust-NExT may have relevance to other spacecraft missions.
Tempel 1↗
Science Benefits of Onboard Spacecraft Navigation
Primitive bodies (asteroids and comets), which have remained relatively unaltered since their formation, are important targets for scientific missions that seek to understand the evolution of the solar system. Often the first step is to fly by these bodies with robotic spacecraft. The key to maximizing data returns from these flybys is to determine the spacecraft trajectory relative to the target body-in short, navigate the spacecraft- with sufficient accuracy so that the target is guaranteed to be in the instruments' field of view. The most powerful navigation data in these scenarios are images taken by the spacecraft of the target against a known star field (onboard astrometry). Traditionally, the relative trajectory of the spacecraft must be estimated hours to days in advance using images collected by the spacecraft. This is because of (1)!the long round-trip light times between the spacecraft and the Earth and (2)!the time needed to downlink and process navigation data on the ground, make decisions based on the result, and build and uplink instrument pointing sequences from the results. The light time and processing time compromise navigation accuracy considerably, because there is not enough time to use more accurate data collected closer to the target-such data are more accurate because the angular capability of the onboard astrometry is essentially constant as the distance to the target decreases, resulting in better "plane-of- sky" knowledge of the target. Excellent examples of these timing limitations are high-speed comet encounters. Comets are difficult to observe up close; their orbits often limit scientists to brief, rapid flybys, and their coma further restricts viewers from seeing the nucleus in any detail, unless they can view the nucleus at close range. Comet nuclei details are typically discernable for much shorter durations than the roundtrip light time to Earth, so robotic spacecraft must be able to perform onboard navigation. This onboard navigation can be accomplished through a self- contained system that by eliminating light time restrictions dramatically improves the relative trajectory knowledge and control and subsequently increases the amount of quality data collected. Flybys are one-time events, so the system's underlying algorithms and software must be extremely robust. The autonomous software must also be able to cope with the unknown size, shape, and orientation of the previously unseen comet nucleus. Furthermore, algorithms must be reliable in the presence of imperfections and/or damage to onboard cameras accrued after many years of deep-space operations. The AutoNav operational flight software packages, developed by scientists at the Jet Propulsion Laboratory (JPL) under contract with NASA, meet all these requirements. They have been directly responsible for the successful encounters on all of NASA's close-up comet-imaging missions (see Figure !1). AutoNav is the only system to date that has autonomously tracked comet nuclei during encounters and performed autonomous interplanetary navigation. AutoNav has enabled five cometary flyby missions (Table!1) residing on four NASA spacecraft provided by three different spacecraft builders. Using this software, missions were able to process a combined total of nearly 1000 images previously unseen by humans. By eliminating the need to navigate spacecraft from Earth, the accuracy gained by AutoNav during flybys compared to ground-based navigation is about 1!order of magnitude in targeting and 2!orders of magnitude in time of flight. These benefits ensure that pointing errors do not compromise data gathered during flybys. In addition, these benefits can be applied to flybys of other solar system objects, flybys at much slower relative velocities, mosaic imaging campaigns, and other proximity activities (e.g., orbiting, hovering, and descent/ascent).
Autonomy↗
Navigation of the EPOXI Spacecraft to Comet Hartley 2
On November 4, 2010, the EPOXI spacecraft flew by the comet Hartley 2, marking the fourth time that a NASA spacecraft successfully captured high resolution images of a cometary nucleus. EPOXI is the extended mission of the Deep Impact mission, which delivered an impactor on comet Tempel-1 on July 4, 2005. EPOXI officially started in September 2007 and eventually took over 3 years of flight time and had 3 Earth gravity assists to achieve the proper encounter conditions. In the process, the mission was redesigned to accommodate a new comet as the target and changes in the trajectory to achieve better imaging conditions at encounter. Challenges in navigation of the spacecraft included precision targeting of several Earth flybys and the comet encounter, uncertainties in determining the ephemeris of the comet relative to the spacecraft, and the high accuracy trajectory knowledge needed to image the comet during the encounter. This paper presents an overview of the navigation process used for the mission.
Deep Impact (DI)↗
Ground Optical Navigation for the Stardust-Next Mission to Comet 9P/TEMPEL1
Ground-based optical navigation (OpNav) using pictures taken by the Naviga-tion camera on the Stardust spacecraft provided the target-relative information needed to design maneuvers during its approach to comet Tempel 1. Hardware problems, limited downlink bandwidth, and changes in the flight profile affected the OpNav picture schedule, sometimes in near-real time. The Stardust naviga-tion camera and attitude control presented challenges. Picture-processing techniques were developed during approach that included background estimation, co-addition, and co-registration. These techniques, along with adaptive picture scheduling, successfully addressed the challenges.
9P/Temple 1↗
Navigating Stardust-NEXT: The Road to Tempel 1
The Stardust-NExT (New Exploration of Tempel) mission, a follow-on to the Stardust prime mission, successfully completed a flyby of comet Tempel-1 on 2/14/11. However there were many challenges along the way in navigating this mission to its successful conclusion, most significantly low propellant margin and detection of the comet in imagery later than anticipated. These challenges and their ramifications forced the navigation team and the project to respond with flexibility and ingenuity. As a result, the resulting flyby at an altitude of 178 km was nearly flawless, accomplishing all its science objectives.
flyby↗
Optical Navigation for the EPOXI Mission
The Deep Impact spacecraft flew by comet Hartley 2 on November 4, 2010 as part of its extended mission called EPOXI. Successful navigation depended critically on the quality and timing of optical navigation data processing, since pictures of the comet provided the most precise comet-relative position of the spacecraft. This paper describes the planning, including the picture timing and pointing; the methods used to determine the center of the comet image in each picture; and the optical navigation results, which provided the necessary information to allow the cameras to accurately target the comet for science imaging at encounter.
small-body mission↗
Small Body Landing Accuracy Using In-Situ Navigation
Spacecraft landings on small bodies (asteroids and comets) can require target accuracies too stringent to be met using ground-based navigation alone, especially if specific landing site requirements must be met for safety or to meet science goals. In-situ optical observations coupled with onboard navigation processing can meet the tighter accuracy requirements to enable such missions. Recent developments in deep space navigation capability include a self-contained autonomous navigation system (used in flight on three missions) and a landmark tracking system (used experimentally on the Japanese Hayabusa mission). The merging of these two technologies forms a methodology to perform autonomous onboard navigation around small bodies. This paper presents an overview of these systems, as well as the results from Monte Carlo studies to quantify the achievable landing accuracies by using these methods. Sensitivity of the results to variations in spacecraft maneuver execution error, attitude control accuracy and unmodeled forces are examined. Cases for two bodies, a small asteroid and on a mid-size comet, are presented.
in-situ navigation↗
Trajectory Reconstruction of the ST-9 Sounding Rocket Experiment Using IMU and Landmark Data
This paper presents trajectory reconstruction of the ST-9 sounding rocket experiment using the onboard IMU data and descent imagery. The raw IMU accelerometer measurements are first converted into inertial acceleration and then used in trajectory integration. The descent images are pre-processed using a map-matching algorithm and unique landmarks for each image are created. Using the converted IMU data and descent images, the result from dead-reckoning and the kinematic-fix approaches are first compared with the GPS measurements. Then, both the IMU data and landmarks are processed together using a batch least-squares filter and the position, velocity, stochastic acceleration, and camera orientation of each image are estimated. The reconstructed trajectory is compared with the GPS data and the corresponding formal uncertainties are presented. The result shows that IMU data and descent images processed with a batch filter algorithm provide the trajectory accuracy required for pin-point landing.
Park, Ryan S.↗
Estimating Small-Body Gravity Field from Shape Model and Navigation Data
This paper presents a method to model the external gravity field and to estimate the internal density variation of a small-body. We first discuss the modeling problem, where we assume the polyhedral shape and internal density distribution are given, and model the body interior using finite elements definitions, such as cubes and spheres. The gravitational attractions computed from these approaches are compared with the true uniform-density polyhedral attraction and the level of accuracies are presented. We then discuss the inverse problem where we assume the body shape, radiometric measurements, and a priori density constraints are given, and estimate the internal density variation by estimating the density of each finite element. The result shows that the accuracy of the estimated density variation can be significantly improved depending on the orbit altitude, finite-element resolution, and measurement accuracy.
Covariance Analysis↗
The Next 25 Years of Deep Space Navigation
This slide presentation reviews the missions that will be flown into deep space in the next 25 years, the navigational challenges for these missions, and the strategies that will be used to overcome these challenges. The challenges include: (1) an incresed need for autonomous navigation, (2) an increased use of in-situ and optical navigation, (3) an increased use of low-thrust propulsion, (4) an increased need for higher accuracy in guidance, navigation, and control, and an increased need for integration between flight path and attitude control. The enabling strategies that are planned for use are: (1) Advance Radio-Metric Tracking Capabilities, (2) Expand the Use of Optical Navigation, (3) Develop General-Purpose Autonomous Navigation Capabilities, (4) Improve Frequency and Timing Systems, and (5) Develop In-situ Tracking Infrastructure. Future trends that are being developed are Optical and Autonomous Navigation
optical navigation↗
The Next 25 Years of Deep Space Navigation
This paper reviews the most probable set of NASA deep space missions that will be launched in the next twenty-five years, discusses the navigational challenges that will confront them, and outlines the most probable solutions to these challenges.
communications networks↗
AutoNav Mark3: Engineering the Next Generation of Autonomous Onboard Navigation and Guidance
The success of JPL's AutoNav system at comet Tempel-1 on July 4, 2005, demonstrated the power of autonomous navigation technology for the Deep Impact Mission. This software is being planned for use as the onboard navigation, tracking and rendezvous system for a Mars Sample Return Mission technology demonstration, and several mission proposals are evaluating its use for rendezvous with, and landing on asteroids. Before this however, extensive re-engineering of AutoNav will take place. This paper describes the AutoNav systems-engineering effort in several areas: extending the capabilities, improving operability, utilizing new hardware elements, and demonstrating the new possibilities of AutoNav in simulations.
navigation↗