V3351296_-_SOW.docx

STATEMENT OF WORK

Autonomous Unmanned systems Expert Support

1.0 BACKGROUND

The U.S. Department of Transportation, John A. Volpe National Transportation Systems Center (Volpe Center) is researching and assessing the application of artificial intelligence with autonomous vehicle technology. The Volpe Center, in conjunction with other Federal partners, are working collective within the National Capital Region. The dynamic and rapid advancement in and ad hoc application of machine learning in vehicle platforms (i.e. cars, trucks, boats, unmanned aerial systems (UAS) requires the Volpe Center and Federal partners to quickly assess usage that individual may attempt that could present unsafe conditions on our roadways, waterways, and airspace.

2.0 SCOPE & OBJECTIVE

The Contractor will provide a single location within 30 minutes/25 miles of Arlington, VA with indoor laboratory and storage space to allow Government engineers to develop/repair/fabricate remote controlled unmanned and autonomous vehicle platforms (aerial systems [UAS] and car), and exterior testing area for these platforms. Concurrently, the Contractor will design, fabricate, and demonstrate a mini-quadcopter flying outside autonomously around objects and landing on a designated locations using only commercial off-the-shelf (COTS) components and public available open-source code.

3.0 REQUIREMENTS (TASKS)

This project has four primary tasks to be accomplished at the same facility within 30 minutes/25 miles of Arlington, VA: (1) provide a shared work area for Government engineers and their on-site support staff to develop/repair/fabricate remote controlled and autonomous vehicle platforms (UAS and car); (2) interior secure storage for up to 30 UAS and exterior storage of 1 electric/hybrid car; (3) exterior, limited access, open area to safely test UAS and electric/hybrid car operations either under remote control or operating semi- to fully autonomously and (4) design, develop, and demonstrate the application of publically available, open-source algorithms to implement machine learning to recognize objects and autonomously fly a quadcopter.

Task 1 Indoor Laboratory Access for Government Engineers

The Contractor shall provide at an indoor laboratory work area for Government engineers and their support contractors. No more than 2 individuals will be use the provided interior work area on a normal day. This area will be used approximately 60% of the time over the duration of the contract. These engineers will use this space for the development and application of artificial intelligence/machine learning to autonomous vehicle navigation and operation systems, and to maintain/repair the Government vehicle platforms.

For Government engineers using the laboratory, the Contractor shall provide:

· Up to two work bench areas with standard electronic test equipment (computer/software, power supplies, multimeter, soldering equipment, and similar material) used in the maintenance/repair/building of autonomous vehicles

· 3D printer capable of producing the type and quantity of parts to replace small components of autonomous vehicles

· Laser milling machine

The laboratory can be used by Contractor staff engaged in similar innovative unmanned vehicle design, fabrication and programming activities.

Task 2 Storage of Government Test Vehicles

The Contractor shall provide secure interior storage space for up to 30 Class 1 or Class 2 UAS and exterior, protected storage of 1 electric/hybrid car which can be accessed the Government engineers and their support contractors working at the Laboratory or exterior testing facility. The space should be in close proximity to the laboratory space and exterior testing facility. The storage area will be used 100% of the time over the duration of the contract.

Task 3 Exterior Testing Facility

The Contractor shall have an exterior facility co-located with the storage and laboratory facility where autonomous aerial platforms and cars can be safely and legally operated for testing by the Government engineers and their support contractors. The minimum surface area is 360 feet x 160 feet with no ground or overhead obstructions that will damages the aerial or vehicle platforms during operations. The ability to simultaneously operate multiple platforms, without causing interference to the other operator or equipment, is permissible. The area will be used as needed, but no more than 50% of the time over the duration of the contract.

The Contractor shall provide the Government documentation as to the maximum allowed altitude and distances aerial platforms can be operated; and the maximum over-ground distance for cars.

If there are any restriction on the use and operations of remote controlled or semi-autonomous platforms, the Contractor shall provide that information prior to contract award. The testing area can be used by Contractor staff engaged in similar innovative unmanned vehicle design, fabrication and programming activities.

Task 4 Artificial Intelligence/Machine Learning Application to Autonomous Mini-Quadcopter Operation

The Contractor will design, develop, and demonstrate the application of publically available, open-source algorithms to implement machine learning to recognize objects and autonomously fly a quadcopter. The task will focus on agile development of portable open source training codes and programs. The software used shall be open-source (preferably written in OpenCV, Python or TensorFlow) and shall contain no proprietary code, or any pieces of code that has been previously used on or for a US Government funded program that the Contractor or their staff were involved. If at any time this requirement is violated, it will result in termination and non-payment of the contract.

The Contractor will determine what sensor technology and graphic processing unit is required to achieve these objectives and incorporate them onto a quadcopter so to minimize performance problems with the quadcopter. The quadcopter operation should be able to achieve near maximum altitude, speed, and agility as possible when operating with maximum payload. The quad-copter will have the ability to fly autonomously, with the ability to be trained to recognize objects to be avoided and to land on. The training shall allow the objects to be moved to ad hoc locations with minimal performance impact.

The quad-copter used by the Contractor may be commercially available, excluding Department of Defense prohibited systems, or the Contractor may design and build a custom system. The quadcopter and other sensors will be built using COTS resources which are unrestricted to obtain. Within 30 days of task initiation, the Contractor shall provide the Government performance parameters of the smallest quad-copter feasible, such as a mini-quad-copter. If a commercially available quadcopter is used, the system shall be capable of operating near its maximum performance parameters. The Contractor shall provide the Government the performance parameters and physical characteristics of the quadcopter for approval prior to proceeding with the acquisition and construction of the systems.

The Contractor shall design and implement a method to safely and immediately terminate the quad-copter’s operation while it is operating autonomously if the Contractor or Government feel an unsafe condition exists.

The Contractor shall provide on-site project management, monitoring the safety and security of the Government equipment and work area, and reporting monthly on the progress made for each activity in the past calendar month, and planned activities for the forthcoming 2 months. The fiscal reporting, presented in tabular and corresponding line graphs, the planned expenditures per month and cumulative expenditures vs. actual monthly and cumulative expenditures shall be part of the monthly status report. The contractor shall provide the Government a schedule within 30 days detailing the machine learning development and quadcopter design/integration. The schedule shall include periodic reviews which will be held at the contractor facility by the Government.

The Contractor shall provide a complete technical report of the autonomous UAS development with documentation and manuals on how to port the open source software. The report shall be provided electronically, as well as full software coding documentation and developed code to allow Government engineers to continue training the system on objects to recognize.

4.0 DELIVERABLES and MILESTONES

The Contractor will provided the following items within the period of time specified.

5.0 PERIOD OF PERFORMANCE

Performance period is for 1 year from contract award, with a 1-year option period.