Search NASA⌕ Search

Engineering topics

Ono, Masahiro

Publications and source records attributed to Ono, Masahiro.

At least 19 records
All sourcesNASA NTRSDOE OSTI
NASA NTRS2022

Multi-Mission Terrain Classifier for Safe Rover Navigation and Automated Science

We previously presented Soil Property and Object Classification (SPOC), a machine learning-based terrain classifier for Mars rovers, for automatically segmenting rover images by its surface type such as sand and bedrock. This paper presents a number of practical improvements to pave the way for potential future onboard deployment. First, we achieved 97.0% overall pixel accuracy, evaluated against the classification generated by human experts on images from Mars Science Laboratory (MSL) missions. The substantial increase in accuracy was primarily enabled by the sheer volume of data used for training; we created a new large-scale dataset of Martian terrain labels, namely AI4Mars, which contains more than 400k labels contributed by citizen scientists for 50k images taken by the Mars Exploration Rovers (MER) and Mars Science Laboratory (MSL) rover. Second, we demonstrated that SPOC can quickly adapt to a new mission landed on a previously unseen site. Specifically, we pretrained a model with MER and MSL data from the AI4Mars dataset and then adapted to the Mars 2020 Rover (M2020) by feeding a small volume of data between Sol 0 and 157; the adapted model was tested on Sol 200-203 and resulted in 84.2% overall pixel accuracy and 93.4% reliability (recall) for detecting sand, the most concerning class for rover’s traversability. Third, we found that pretraining can substantially mitigate the decline of accuracy over time. We showed that the performance of a SPOC model pretrained with the ImageNet dataset and then trained by MSL images only up to Sol 390 remains comparable to a model trained by images up to Sol 1689 on the test data after Sol 1689. Fourth, we reimplemented SPOC with a light-weight convolutional neural network (CNN), MobileNetV2, which typically runs within tens of milliseconds (ms) on mobile processors such as Qualcomm’s Snapdragon. Finally, we released the AI4Mars dataset to the public to encourage open innovation.

Ono, Masahiro↗
NASA NTRS2021

Machine Learning Based Path Planning for Improved Rover Navigation

Enhanced AutoNav (ENav), the baseline surface navigation software for NASA’s Perseverance rover, sorts a list of candidate paths for the rover to traverse, then uses the Approximate Clearance Evaluation (ACE) algorithm to evaluate whether the most highly ranked paths are safe. ACE is crucial for maintaining the safety of the rover, but is computationally expensive. If the most promising candidates in the list of paths are all found to be infeasible, ENav must continue to search the list and run time-consuming ACE evaluations until a feasible path is found. In this paper, we present two heuristics that, given a terrain heightmap around the rover, produce cost estimates that more effectively rank the candidate paths before ACE evaluation. The first heuristic uses Sobel operators and convolution to incorporate the cost of traversing high-gradient terrain. The second heuristic uses a machine learning (ML) model to predict areas that will be deemed untraversable by ACE. We used physics simulations to collect training data for the ML model and to run Monte Carlo trials to quantify navigation performance across a variety of terrains with various slopes and rock distributions. Compared to ENav's baseline performance, integrating the heuristics can lead to a significant reduction in ACE evaluations and average computation time per planning cycle, increase path efficiency, and maintain or improve the rate of successful traverses. This strategy of targeting specific bottlenecks with ML while maintaining the original ACE safety checks provides an example of how ML can be infused into planetary science missions and other safety-critical software.

Yue, Yisong↗
NASA NTRS2020

A ROS-based Simulator for Testing the Enhanced Autonomous Navigation of the Mars 2020 Rover

In order to achieve the ambitious objectives of the Mars 2020 (M2020) mission, in particular the ability to autonomously traverse more challenging terrains more efficiently, new surface mobility software was developed for Enhanced Navigation (ENav). That decision was made early in the project, before most of the new surface flight software (FSW) existed, which created a need for a separate framework where the new navigation algorithms could be quickly prototyped and tested, before more realistic FSW-based testbeds became available. The JPL robotics team chose the Robot Operating System [1] (ROS) as the environment in which to test the new ENav algorithms. This made it possible to write the algorithms in the C language required by the FSW, so they could be directly ported over to the flight module later on, while leveraging all the C++ libraries and tools provided by ROS for simulation and testing. The ENav algorithms were developed as a separate C library, and stubs were used to replace any FSW-specific code, such as Event Reporting (EVRs) and data products (DPs). A ROS simulator was developed to generate a rich set of varied 3D terrains representative of the candidate Mars landing sites and simulate the physics of the rover motion, the point cloud perceived by the rover’s stereo vision system, and the new thinking-while-driving (TWD) navigation logic which directs the rover to drive autonomously to user-specified waypoints. To simulate the rover motion and perception, a ROS node was developed that uses a software library called HyperDrive Sim (HDSim), which is a wrapper for the Rover Sequencing and Visualization Program [2] (RSVP). That library provides roverterrain settling, realistic slip modelling, and camera rendering capability based on the rover’s NavCam machine vision models. To simulate the navigation logic, a ROS node was created that initializes and runs the ENav algorithms in a way that mimics the FSW execution, while also providing the capability to load and replay data products, including re-running the recorded inputs through the ENav algorithms for testing. An engineering Graphical User Interface (GUI) was also developed to visualize various elements, such as the rover pose during the drive, the simulated and perceived terrain, the selected local and global paths to the goal, the evaluated candidate paths and the reasons why they were rejected, the keep-in and keep-out zones (KIOZs), etc. Finally, an advanced Monte Carlo (MC) framework that can run many simulations in parallel on the Cloud and automatically generate reports that capture the key ENav performance metrics was developed to evaluate the system in a statisticallymeaningful way. This paper provides an overview of the ROSbased simulator used for testing the M2020 ENav algorithms.

Toupet, Olivier↗
NASA NTRS2018

Comet Hitchhiker: NIAC Phase 1 Final Report

Summary of Activities-Developed the Comet Hitchhiker concept, which is to hitch rides on small bodies (asteroids and comets) using a tethered spacecraft. (Section 2)-Identified five scientifically important missions that would be enabled or significantly benefited by the Comet Hitchhiker concept.The five mission concepts are: KBO rendezvous, Centaur rendezvous, Trojan rendezvous, Damocloid rendezvous, and Main asteroid belt tour to rendezvous with multiple (10) objects. (Section 3)-Derived the Space Hitchhike Equation, or "the rocket equation for hitchhiker", which relates the specific strength of tether, mass ratio, and V. (Section 4.1)-Performed in-depth feasibility analysis of the critical components of the concept through: Finite-element simulations of tether and spacecraft dynamics, as shown in Figure 1 (Section 4.4); Supercomputer simulations of the hypervelocity impact of harpoon on a small body, as shown in Figure 2. (Section 6)-Performed public outreach activities including the collaboration with a concept artist of the Museum of Science Fiction, exposure to media, and public presentations. (Section 8)

Deep Space↗
NASA NTRS2018

Enceladus Vent Explorer Concept

Our concept, Enceladus Vent Explorer (EVE), is a robotic pathfinder mission to enter these doors. EVE's goals are to descend into erupting conduits up to ~2 km deep, characterize the unknown interior structure of the vent-conduit system, assess the accessibility to the subsurface ocean through the vent-conduit system, potentially reach the liquid interface, and perform astrobiology and volcanology observations in the vent-conduit system. EVE sends two types of modules: Surface Module (SM) and Descent Module (DM). SM is a lander that stays on the surface, while tens of small (~3 kg, 10 cm in width and 30 cm in length) DMs separate from SM, move to a vent, and descend into it. DMs rely on a power and communication link provided by SM through a cable. As the payload volume of DM is extremely limited, each DM can carry only a single miniaturized instrument. This limitation is complemented by heterogeneity. There are several types of DMs, all of which share the common mobility system but carry different instruments. For example, a "scout DM" creates a 3-D map of the geyser system with its stereo cameras and structured light. A "sample return DM" collects particles and ice cores in the vent and deliver them to the mass spectrometer in the SM. An "in-situ science DM" carries science instruments, such as a microscopic imager and a microfluidics chip for biosignature detection. DMs are sent either sequentially or in parallel.

Enceladus↗
NASA NTRS2017

Mars 2020 Site-Specific Mission Performance Analysis: 1. Science exploration and sol type modeling

We developed a scenario-based mission performance model for the Mars 2020 Rover to help identify project-wide productivity and operability challenges and opportunities. We have modeled science exploration combining high-fidelity resource modeling, science strategies, and modular science activity scenarios. This is proving to be an effective way of incorporating science activities into an engineering-based performance model.

Ono, Masahiro↗
NASA NTRS2017

Mars 2020 Site-Specific Mission Performance Analysis: Part 2. Surface Traversability

The Mars 2020 Rover Mission (M2020) is characterized by long-range traverses between scientific Regions Of Interest (ROIs), as well as the demanding requirement on the distance and time for the inter-ROI traverses. As a result, surface traversability is one of the major driving factors for the landing site selection of M2020. With the newly developed Mars Terrain Traversability analysis Tools (MTTT), we performed traversability analysis of the eight candidate landing sites with an unprecedented granularity. This paper describes the MTTT analysis capabilities, as well as how the MTTT capabilities were used to down-select from eight to three candidate landing sites for further evaluation.

Milkovich, Sarah↗
NASA NTRS2017

Mars 2020 Surface Mission Performance Modeling: Part 3. Mission Performance Modeling Approach and Results

We have developed a comprehensive model of the Mars 2020 rover surface mission used to identify and evaluate mission productivity and operability challenges and opportunities. The surface mission performance model combines many aspects of rover design, operations approach, system behaviors, and mission constraints into an integrated model of surface mission performance. The surface mission model applies monte-carlo methodologies in order to investigate the key sources of mission performance variability and their effects with respect to overall mission performance. Our modeling efforts aims to investigate and inform engineering design and development efforts, as well as the landing site selection process led by the Mars science community. We believe that an integrated mission performance model such as this is important to aligning engineering and science efforts across the project.

Ono, Masahiro↗