SOURCES SOUGHT
A -- Position, Navigation, and Timing (PNT) Technology Readiness for Safe Automated Vehicle (AV) Operations
- Notice Date
- 3/18/2020 3:06:47 AM
- Notice Type
- Sources Sought
- NAICS
- 541715
— Research and Development in the Physical, Engineering, and Life Sciences (except Nanotechnology and Biotechnology)
- Contracting Office
- 6913G6 VOLPE NATL. TRANS. SYS CNTR CAMBRIDGE MA 02142 USA
- ZIP Code
- 02142
- Solicitation Number
- 6913G6-20-SS-00001
- Response Due
- 4/2/2020 12:00:00 PM
- Archive Date
- 04/17/2020
- Point of Contact
- Roland Regan
- E-Mail Address
-
roland.regan@dot.gov
(roland.regan@dot.gov)
- Description
- �THIS IS NOT A REQUEST FOR COMPETITVE PROPOSALS: This is not a solicitation or a request for proposal and no contract will be awarded from this notification. No solicitation or specifications are available at this time. Please monitor SAM.gov, this sources sought will not be linked to a solicitation. Description of Requirement: The Volpe National Transportation Systems Center (Volpe Center) is looking to develop a robust EKF sensor fusion algorithm (in MATLAB) as a mean to provide a baseline assessment of the performance enhancements (including availability, accuracy, and integrity) that can be achieved by customizing a PNT sensor suite in comparison to each individual sensor performance. The assessment is to be carried out for a number of scenarios representative of various autonomous vehicle (AV) operating environments. One of the primary application of such a fusion algorithm is to assess the enhancements of and trade-offs between availability, accuracy, and integrity that can be achieved within the constraints of a representative scenario for a class of realistic scenarios. A scenario is defined by its attributes. Examples of scenario attributes are: Size of autonomous region (highway-subsection, predefined area within a city, the whole city etc...) Route geometry, range of dynamic variables, and duration Urbanization level (open/rural areas, light urban, and urban canyons) and foliage levels as well as other GNSS challenging conditions Visibility and precipitation condition that present a challenge to Lidar and vision systems Assumed cooperation level from surrounding vehicles and infrastructure within the region of autonomy (e.g. smart highway/city with or without V2X communication throughout the autonomous region. Fully cooperative or mixed environment) Landmarks type and density in the autonomous region Different characteristics of the surrounding dynamics: Fully static (buildings, signs, poles, and trees) Static and quasi-static (parked and stopped vehicles, bins, and movable objects)� Static, quasi-static, and dynamic where moving vehicles and pedestrians are also present. Additional application is the resiliency enhancements offered by such filter for various fault modes. The following are example of fault modes of interest. No Fault Spoofed modes: At least one predefined spoofing attack on each of the sensors on a subset of the scenario duration when applicable (for example spoofing is likely inapplicable for an INS system) Jammed modes: Predefined subset of the scenario duration when jamming is applied for each sensor, at least one at a time, when applicable. The dynamic State Space (SS) model part of the EKF should be representative of an AV based on 2D linear and yaw AV dynamics with user modifiable parameters at a minimum. Although not required, having an additional EKF flavor based on an SS model for 6D UAV dynamics as well is a plus. The fusion algorithm platform should be able to accommodate at least the following 4 sensor inputs: GNSS, INS, Lidar, and camera-based localization sensors. The EKF should employ tight GNSS/INS integration. It should be able to provide prediction and update of the states and covariance matrices even if one or more of the 4 sensors are not present. It should be able to use GNSS information as part of the fusion algorithm even when less than 4 satellites are visible. The fusion algorithm should implement smart outlier data rejection and limited spoofing data detection and exclusion algorithms. Fusion algorithm implementation should provide coasted values of all states along with the propagated covariance matrices for time steps when all measurement inputs are no present. At a minimum, the Lidar and vision systems data should be presented to the EKF after it has been processed into position and/or velocity data along with the measurement quality outputs expected to be available from such sensors. For such an implementation, the organization developing the filter should have the ability to produce position and velocity data from these sensors via modeling, simulation, and/or data measurements for the scenarios considered. Such data should conform to the expected error distribution that each sensor is expected to produce in such scenarios. The error distribution should include outliers (such as the errors due to the probability of false classification of landmarks associated with particular Lidar measurements or vision system) in a particular scenario or class of scenarios.� For example, the error distribution (including outliers due to the probability of misclassification and misdetection) of Lidar-based position measurements in an environment with densely deployed landmarks or mapped features in a feature rich environment and in the presence of other moving vehicles can be developed via available measurements, models derived from measurements available to the organization developing the EKF, simulation of Lidar measurements in similar or representative environment that are accessible to the organization developing the EKF, or other means. For the case of GPS and INS, the data (measured/and or simulated) for a particular scenario should include all the measurements needed for the tight integration implementation of the EKF. Similarly, the error distributions for these measurements should conform with each representative scenario under consideration including outliers for GNSS measurements (for example large errors due to inability to reject some multipath, poor DOP, and or cycle slips due to signal attenuation and intermittent multipath in a subset of routes within a scenario). While such errors are a function of the antenna(e), receiver architecture, augmentation data, and signal processing or GNSS signals, the error distributions and/or data collection for error characterization should be representative of GNSS receiver grades planned for future and emerging AV systems. As such outreach to the industry is an important part of this work.� The fusion algorithm should also implement data screening and outlier rejection algorithms prior to them contaminating the filter updates. This screening should make use of information from the states and measurement covariance matrices and other information that can be inferred during the execution of each representative scenario. The filter should be able to take in sensor specification information as an input to evaluate the same scenario with higher and lower grade sensors. The work also involves defining the AV representative scenarios jointly with Volpe, generating input data representative of these scenarios via simulation and/or test as well as using publicly and other available and relevant datasets. To be clear, it is the responsibility of the organization developing the EKF to find and/or generate position data (modeled or measured) for each sensor with an error distribution conforming to the attributes of each of the representative realistic scenarios under consideration and sensor specifications. The data (simulated, collected, and/or modeled) for each of the sensors under each scenario (or a scenario representative of a class of scenarios) should be large enough to evaluate integrity values that are at least up to 10-4 probability of hazardous misleading information (HMI). Scenarios for which the EKF-fused position output exhibits an integrity level better than 10-4 HMI, qualitative assessment and/or analytical extrapolation should be performed by the organization to infer the limit of the improved integrity the fusion algorithm is likely to meet down to 10-7 HMI probability. To perform these tasks, the Volpe Center will assemble and manage a Review Panel consisting of three (3) engineers with PNT expertise and understanding of automated vehicle sensing localization technology. Contractor personnel for these seven (7) tasks shall consist of engineers experienced with Matlab, PNT, AV sensing and localization technology, data processing and fusion algorithms, safety assessment approaches. Such personnel must also possess good communication skills and capable of providing updates at periodic group meetings. Such personnel should be able to act as integral member(s) of the Volpe team working on the safety assessment of AV PNT sensor suites for emerging and future AV applications. In order to best provide the needed quality support, the Volpe Center anticipates assistance in the following seven (7) task areas. Work with Volpe to jointly define the AV representative scenarios. � Provide results for fusion filter for each of the selected scenarios including the tradeoff analysis between availability, integrity, accuracy, as well as resiliency metrics under fault conditions. � Analytical development of the fusion algorithm to accommodate at least the four (4) sensor inputs: GNSS, INS, Lidar, and camera-based localization sensors. Fusion algorithm should also implement data screening and outlier rejection algorithms prior to them contaminating filter updates. � Real and/or simulated measurements and truth data for each of the considered scenarios for individual sensors. The data files from the vision and Lidar systems (collected and simulated) should be provided in both raw form (if applicable) and in the form of processed localization data and measurement uncertainty bounds (or measurement covariance). The processed form should be in a format directly useable by the Matlab EKF implementation. � Extrapolation analysis of integrity values for the cases when probability of HMI is too small to be measurable with the available data set. � Documented and user friendly Matlab modules (including main/wrapper function(s) and configuration files with or without GUI interface(s)). Contribute writing and editing to the appropriate sections or chapters of Volpe/DOT report deliverable reports. � All interested firms shall submit a written capabilities statement (maximum eight pages) addressing all of the above seven (7) tasks by providing clear and convincing detailed documented evidence of the firms' capability to provide the required seven task areas as follows: Identify past and current professional experience which provides clear and convincing documented evidence demonstrating that your firm is qualified to perform work under all seven (7) task areas identified above. Without the required documentation your response will be considered non-responsive and will not be considered.� Please include points of contact and telephone numbers for all referenced projects for verification. � Does your firm currently qualify to compete under NAICS 541715? If yes, and assuming question #1 above has provided clear and convincing documented evidence of your capabilities, would your firm compete for this requirement if a solicitation was issued? � Does your firm currently have the capability to meet all seven (7) tasks no matter where or when they may occur? (If yes, please address in detail how you would meet each task through current capabilities and/or national locations of current offices?). � Is your firm currently registered in System for Award Management (SAM)?� If yes, please provide both your SAM identification number and DUNS number, and ensure you�re Reps and Certs are current or up to date in SAM. https://www.sam.gov/portal/pub!ic/SAM/. If not, you are required to be registered in SAM to do any business with the government. Do you have an approved DCAA accounting system? If yes, when was this written approval issued by the DCAA? Otherwise if not approved in writing by the DCAA, please so state. � Your capabilities statement shall address all five (5) questions and shall not exceed eight (8) type written single pages using 11 font size and Times New Roman. Failure to provide the requested detail addressing all seven tasks and responses to all five (5) questions within the eight (8) single page maximum will result in your firm not being considered responsive. � AGAIN, THIS IS NOT A REQUEST FOR COMPETITIVE PROPOSALS.� Interested firms must submit a written capability statement (maximum six pages) to the cited named point of contact providing clear and convincing evidence of the firms' capability to provide the required services.� Written capability statements must be submitted within fifteen (15) calendar days from the date of publication of this synopsis and presented in the format cited above.� Responses received after fifteen (15) calendar days or without the required complete and detailed documentation as required will be considered non-responsive and will not be considered.� Such documentation will be utilized solely for the purpose of determining whether or not to conduct this procurement on a competitive basis.� The Government will not pay for any documentation provided in response to this synopsis and documentation received will not be returned to the sender.� Such documentation will be utilized solely for the purpose of determining whether or not to conduct this procurement on a competitive basis. The Government will not accept any face-to-face meetings, phone calls, skyping, and/or face-timing regarding this or any subject-matter related to this requirement from any 3rd party source or contractor. A determination by the Government to compete or not to compete this requirement on a full and open competitive basis will be based upon responses received to the synopsis and is solely within the discretion of the Government.� � Statements of Capability should be sent no later than March 18, 2020 at 3:00 PM ET to: Roland Regan, Volpe National Transportation Systems Center, 55 Broadway, Cambridge, MA 02142 or by email to Roland.Regan@dot.gov.
- Web Link
-
SAM.gov Permalink
(https://beta.sam.gov/opp/5a169f967b0743dd87de31b7370422f7/view)
- Place of Performance
- Address: USA
- Country: USA
- Country: USA
- Record
- SN05593670-F 20200320/200318230225 (samdaily.us)
- Source
-
SAM.gov Link to This Notice
(may not be valid after Archive Date)
| FSG Index | This Issue's Index | Today's SAM Daily Index Page |