Lisa Randall
My feedback
38 results found
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment -
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
The FCPI technical parameters and system design requirements require the integration of multimodal data sources (e.g., those representing "incidents," "weather," "work zones," and "truck parking availability”) to forecast travel time reliability, congestion probability, and truck delay at a corridor level. Predicting the specific precursor for individual vehicle incidents (e.g., overheating vehicles, individual vehicle disablements, or specific cargo issues) is outside the primary focus of the FCPI project. Solutions focused on predicting individual commercial vehicle failures before they occur are considered in the 26-OS1: Predictive Safety Analytics for Commercial Transport Modernization project.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Phase I proposals should define a clear, scalable pathway documenting how the proposed predictive outputs from the AI model will eventually integrate with existing agency systems and workflows. Specifically, the Phase I Conceptual Architecture and System Design should include information to demonstrate how the decision-support solution will be compatible with common enterprise platforms currently used by many public sector agencies (e.g., ArcGIS, Power BI, or Tableau.)
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Truck parking availability, utilization, rest-area saturation, and related conditions can be considered relevant predictive variables. The solicitation notes that "truck parking availability (based on historic parking data and predictive algorithms to forecast parking availability by time of day and day of week)" is one of the sources that can be integrated into the proposed predictive AI model. Identifying truck parking conditions also supports the later commercialization goals of the FCPI project.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Phase I applicants may use synthetic freight datasets or simulated corridor conditions, and Phase I deliverables should demonstrate how the proposed predictive AI model is capable of integrating and analyzing "combined real-world and synthetic freight datasets". Note that applicants are not expected to integrate live operational data during the Phase I period of performance. The primary intent of Phase I is to test data interoperability and demonstrate the feasibility of a predictive model.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
: A Phase I proof of concept utilizing publicly available datasets for the initial model, while planning for private data partnerships in Phase II would be responsive. The primary intent of Phase I is to test data interoperability and demonstrate the technical feasibility of a proposed predictive AI model using combined real-world and synthetic datasets. The Phase I Data Inventory and Integration Plan requires an assessment of public and private data sources and an integration strategy that defines the access methods and availability needed for Phase II implementation.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
: A Phase I proof of concept utilizing publicly available datasets for the initial model, while planning for private data partnerships in Phase II would be responsive. The primary intent of Phase I is to test data interoperability and demonstrate the technical feasibility of a proposed predictive AI model using combined real-world and synthetic datasets. The Phase I Data Inventory and Integration Plan requires an assessment of public and private data sources and an integration strategy that defines the access methods and availability needed for Phase II implementation.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
USDOT is not able to consider adding a new topic or expanding a current topic to the forthcoming FY26 Solicitation. Please note that there is a commercial vehicle safety project (26-OS1 Predictive Safety Analytics for Commercial Transport Modernization) that may be worth reviewing given the operational events noted in the question.
-
3 votes1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
The FCPI project has specific cybersecurity and data protection expectations for handling sensitive and proprietary freight data. To protect proprietary business, occupant or vehicle information, the project emphasizes the use of federated learning methods and federated data exchange protocols. This approach allows for model training and validation across multiple jurisdictions while minimizing the need to transfer raw, sensitive data. Proprietary data should be aggregated, anonymized, and de-identified at the corridor or segment level. Proposed solutions are expected to utilize end-to-end encryption, secure APIs, and credential-based, role-limited access.
-
2 votes1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Aggregated, anonymized, or proprietary commercial freight datasets are acceptable and encouraged for Phase I model development. Individual vehicle trajectories and proprietary business information should never be exposed. Applicants can apply privacy-preserving techniques (e.g., federated learning methods, etc.) to protect proprietary information during model training and validation. Documentation and assessment of data (including data limitations) is expected to be included in the Phase I “Data Inventory and Integration Plan”.
Lisa Randall
supported this idea
·
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
USDOT does not mandate specific data schemas or geospatial formats for Phase I, but there are interoperability expectations. Instead of prescribing specific interoperability schemas, Phase I places the responsibility on the offeror to define them. Phase I deliverables should include a "Data Inventory and Integration Plan" that details specific "data integration strategy defining access methods, availability, and preprocessing steps," alongside a "Conceptual Architecture and System Design" that shows how these streams will be integrated. The proposed Phase I schemas and APIs should clearly demonstrate a scalable path to meeting the broader compatibility requirements outlined for Phase II.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
USDOT does not mandate a specific preferred type or scale of freight corridor for the Phase I proof-of-concept. For Phase I, the primary requirement is to develop a preliminary model for a "selected freight corridor or segment" to demonstrate the technical feasibility of a proposed predictive AI model and to achieve the baseline prediction accuracy of at least 80 percent.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
USDOT does not mandate a specific baseline methodology or a single preferred definition for "freight delay." Applicants can define “delay” based on travel time increases, route deviation, dwell times or another proposed metric, but the approach must ultimately demonstrate that it is capable of achieving a "baseline corridor-level prediction accuracy of at least 80% for short-term (30–60 minute) congestion and delay forecasts". Applicants should also define and provide a rationale for the selected metrics, and their evaluation criteria (e.g., accuracy, timeliness, completeness)" necessary to conduct preliminary benchmarking for the proposed AI model.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
A private-sector freight operator, logistics provider, or transportation management environment may serve as the primary Phase I use case. The Freight Corridor Predictive Intelligence (FCPI) project is intended to investigate capabilities that enhance both public and private freight operational decision-making. However, ensuring that the proposed solution is designed for future integration with public-sector systems is a critical component of a responsive proposal. For the Phase I Commercial and Implementation Readiness Assessment, proposals should state how the Phase II commercialization pathways, partnerships, and integration opportunities with both logistics data providers and public agencies (e.g., state DOTs, local agencies) will be evaluated.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Phase I is a proof of concept on a defined operational segment, freight bottleneck, or intermodal connector; an entire interstate corridor is not expected or required at this stage. The primary Phase I objective is to demonstrate the technical feasibility of the proposed predictive AI model by achieving a baseline prediction accuracy of at least 80%.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Port-linked maritime freight flows connected to a U.S. freight corridor are within the scope of the FCPI project. This project acknowledges that the public infrastructure critical to freight carriers includes "maritime ports" in addition to highways, bridges, and interchanges. Integrating port-linked flows directly aligns with the project's overarching objective to develop an AI-enabled system that fuses "multimodal data sources" to forecast supply chain disruptions and freight bottlenecks. Phase II specifies that field pilots for awarded projects be conducted along 2–3 representative freight corridors that encompass "urban, rural, and intermodal" environments.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Non-AV commercial vehicle telemetry is accepted as a qualifying data input. Connected infrastructure data sources are not limited. Leveraging data from private-sector telematics and logistics data providers (e.g., ATRI, INRIX, Drivewyze, or PrePass) is a primary method to aggregate corridor-level freight movement data. Achieving interoperability with "commercial telematics platforms" and "freight carrier systems" is expected to be a requirement for future awards, scaling and commercialization.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
A Phase I proof of concept may account for cross-border freight variables and border-adjacent congestion, provided the preliminary model remains focused on a selected U.S. freight corridor or segment. Cross-border congestion, customs-related dwell times, and international supply chain disruptions qualify as factors that heavily impact the performance, travel time reliability, and truck delay of border-adjacent U.S. corridors.
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
A benchtop hardware prototype or a hardware-in-the-loop simulation is sufficient to demonstrate technical feasibility for Phase I. Actual field deployment is not required in Phase I. Rather, the primary objective is to demonstrate the technical feasibility and build a proof-of-concept. To meet the Phase I specifications, offerors must develop a "preliminary model for a selected freight corridor or segment" that proves the suggested AI architecture can effectively integrate combined real-world and synthetic datasets. In addition, the Phase I "Conceptual Architecture and System Design" deliverable should outline how predictive modeling, data elements, and hardware components will operate in Phase II deployments
-
1 vote1 comment · U.S. DOT FY 2026 Phase I Pre-Solicitation Q&A » 26-OS2: Freight Corridor Predictive Intelligence · Admin →
An error occurred while saving the comment Lisa Randall
commented
Informal discussions and identifying specific potential partners are sufficient to meet the baseline requirements for Phase I. Demonstrating a higher level of partner commitment during Phase I would help establish a proposal’s overall readiness for Phase II.
Truck parking availability, utilization, rest-area saturation, and related conditions can be considered relevant predictive variables. The solicitation notes that "truck parking availability (based on historic parking data and predictive algorithms to forecast parking availability by time of day and day of week)" is one of the sources that can be integrated into the proposed predictive AI model. Identifying truck parking conditions also supports the later commercialization goals of the FCPI project.