U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A
- To create a new account or to sign in with an existing account please click on “sign in” located at the top right of the page
- A new pop-up window will open to make a selection, you can use an existing Google or Facebook account to access the site or you might select to create a new account.
- If you choose to access UserVoice with an existing Google or Microsoft account, you will be required to enter your login credentials for Google or Microsoft and agree to terms of service
- If you need to create a new account, click "Create an Account" underneath the password box
- In the next screen, type in the e-mail and click "Verify E-mail".
- Navigate to your inbox and click the link within the e-mail to verify. The link will be generated by noreply@trymagic.com
- That link will be live for 20 minutes. Once that time elapses, you will have to request another one
- Navigate back to UserVoice, and you will need to provide your name, a password, and agree to terms of service and storage permissions. Click "Create Account".
- Select the research topic your question is related to from the drop-down menu that appears when you begin typing.
- You may add more detail in the Description box (optional). Do NOT attach any files. The U.S. DOT shall not consider any submitted materials other than questions.
- Enter your e-mail address and password.
- When finished, select “Post idea.” If you registered, you will be able to track responses to your post. Please include only one question or comment per post to help us keep the forum organized.
229 results found
-
Is CARMA / Saxton Lab integration expected for the HIL demo?
For the HIL setup, does the program want to see integration with CARMA / the Saxton Lab, or is it fine to use our own HIL environment? Trying to figure out if CARMA integration is something that actually matters competitively here.
1 vote -
C-V2X as baseline given 5.9 GHz reallocation, or is DSRC still acceptable?
With the 5.9 GHz reallocation and the move to C-V2X, should we be treating C-V2X (PC5/Uu) as the baseline for the HIL demo, or is DSRC still on the table?
1 vote -
Commercial Readiness Assessment — are signed LOIs expected from partners?
For the Commercial and Implementation Readiness Assessment, what level of partner engagement are you expecting? Signed LOIs from state DOTs or MPOs, or is earlier-stage interest enough?
2 votes -
Data integration plan — strategy document or working ingest pipelines?
Is this more of a strategy document, or do you want to see working ingest pipelines for at least some of the data sources by the end?
1 vote -
Digital twin expectation — microsimulation, generative AI, or existing platform integration?
For the digital twin piece — are you thinking more along the lines of physics-based microsimulation (Aimsun, SUMO, Vissim), or a generative AI model trained on real corridor traces? And is the program open to building on top of an existing digital-twin platform, or would you rather see something built from scratch?
1 vote -
How much FL and edge AI must be demonstrated at proof-of-concept stage?
How much of the federated learning and edge AI side does the program want to actually see working at the proof-of-concept stage? We could stand both up at small scale early, or start with a centralized cloud baseline and layer them in later. Curious which the program prefers.
2 votes -
Priority among data sources, and are licensed feeds (NPMRDS, ATRI) government-furnished?
Is there a priority order on the data sources? We're weighing NPMRDS, ATRI, state WIM, FAF, TPIMS, and connected vehicle BSMs, and it'd help to know which ones the program sees as most important. Also, for the licensed ones like NPMRDS and ATRI — are those typically furnished to awardees, or should we plan to budget for them?
1 vote -
Clarification on 80% accuracy target — spatial unit and error metric?
Quick clarification on the 80% accuracy target for 30–60 minute forecasts — is that measured at the segment level, at specific bottlenecks, or across full origin-destination pairs? And what error metric are you using? Whether it's MAPE on travel time vs. classification accuracy on a congestion label changes the modeling approach quite a bit.
1 vote -
Alternate approach to traffic congestion prevention and mitigation
This SBIR solicitation focuses on traffic congestion prevention and mitigation via advisories intended to optimize overall traffic flow. Would SBIR consider funding an alternate approach to traffic congestion prevention and mitigation, which allows all motor vehicles of any type to operate safely in near bumper-to-bumper proximity, thereby significantly increasing the throughput of all roadways? This is achievable without self-driving and with complete ad hock autonomy of each vehicle (i.e., no platoons or preplanning of any type). Fully automated acceleration-deceleration is achievable with adjacent lead-follow vehicle proximity dependencies including speed, braking performance, road conditions, and accommodations to permit ad hock vehicle…
1 vote -
Standards and Compliance Contexts
Does "explainable" imply a particular standard — SHAP/LIME-style local explanations, audit-grade decision logs, regulatory-aligned model cards — or domain-specific explainability tied to DOT compliance contexts?
2 votes -
Trusted Intermediary
Should "Trusted Intermediary" be read as a specific institutional form — federal agency, nonprofit consortium, neutral third party — or is the topic indifferent as long as the technical privacy guarantees hold?
2 votes -
Existing Data Partnerships
Is having a private fleet data partner committed at Phase I a competitive prerequisite, or is securing one part of the Phase II commercialization plan?
5 votes -
Phase I Scope Expectation
What is the expected Phase I demonstration scope on $200K over 6 months — a working prototype validated against real fleet data, or a feasibility framework demonstrated on synthetic / public datasets with Phase II reserved for live integration?
4 votes -
Does the topic favor a specific privacy-preserving approach — federated learning, differential privacy, secure enclaves, or homomorphic encr
Does the topic favor a specific privacy-preserving approach — federated learning, differential privacy, secure enclaves, or homomorphic encryption — or is the choice left to the proposer based on use case?
2 votes -
Is condition-monitoring data from in-vehicle pre-trip inspections — tire tread depth, brake adjustment indicators, defect detection — consid
Is condition-monitoring data from in-vehicle pre-trip inspections — tire tread depth, brake adjustment indicators, defect detection — considered in-scope as "private industry data" for the Trusted Intermediary framework? Or is the topic specifically targeting retrospective operational data such as HOS logs, ELD telematics, and CSA history?
3 votes -
1 vote
-
1 vote
-
1 vote
-
1 vote
-
1 vote