The Road
Must Think
for Itself
Infrastructure that governs intersections at the speed of physics — not the speed of networks.
A full-spectrum procurement document covering system architecture, safety engineering, sensor intelligence, hardware governance, privacy design, Chicago pilot RFP, ROI modeling, federal funding alignment, verification and certification pathway, and competitive displacement analysis.
Roads Are Passive.
Deaths Are Not.
The global transportation network kills more people annually than most armed conflicts. The technology to prevent the majority of these deaths exists today — but it is not deployed at the infrastructure level where it matters most.
Current traffic infrastructure operates on principles established in the 1920s: fixed-interval timers, passive signal hardware, and no awareness of the vehicles, pedestrians, and hazards actively present at any given moment. A traffic light has no idea whether a child is crossing against the signal, whether an ambulance is two blocks away, or whether a vehicle is traveling 60 mph toward a red light it will not stop for.
Americans killed in traffic crashes annually — more than one every 15 minutes, every day
Annual economic cost of traffic crashes in medical expenses, lost productivity, and property damage
Of serious crashes attributable to human choice or error — the majority preventable by infrastructure-level intervention
Global traffic deaths per year — the leading cause of death for people aged 5 to 29 worldwide
These are not inevitable outcomes. They are the predictable result of deploying 21st-century vehicles on 20th-century infrastructure. The intersection — where most fatal collisions occur — remains a passive, timer-governed piece of hardware with no situational awareness. UMIN changes that at the infrastructure level.
Infrastructure That Decides,
Not Merely Responds.
UMIN is a distributed edge AI infrastructure system that transforms road nodes — intersections, corridors, highway segments — into active participants in traffic safety and mobility governance.
At the heart of UMIN is a paradigm shift: the intelligence that prevents collisions and governs intersection behavior must live at the edge — physically, at the pole — not in a cloud data center hundreds of milliseconds away. Life-safety decisions require deterministic, sub-100-millisecond response times. That is not achievable through cloud dependency. It requires hardware-level architecture built for the job.
Each UMIN Smart Pole Unit is an autonomous AI-equipped infrastructure node containing multi-modal sensor arrays, edge compute hardware, V2X communication capability, and an adaptive signal controller — all integrated with a power management system capable of 72-hour autonomous operation during grid outages. The nodes are modular by design, with a 20-year structural lifespan and a 5-year technology upgrade cycle, making UMIN a long-term infrastructure investment rather than a consumer electronics refresh cycle.
What makes UMIN architecturally distinct is not any single sensor or any single AI model — it is the governance structure of how decisions are made, which layer holds authority over which decisions, and what physically cannot happen regardless of software state.
Three Loops.
One Purpose.
UMIN operates three simultaneous closed-loop control tiers, each operating at its natural temporal scale, each holding authority only over decisions appropriate to that scale.
| Loop | Timescale | Functions |
|---|---|---|
| Fast Loop | <100ms | Collision AvoidancePedestrian ExtensionEmergency Pre-emptionNo Cloud Dependency
Hard latency bound enforced. All decisions execute locally at the edge node. |
| Medium Loop | Seconds–Minutes | Adaptive Signal TimingEmergency Corridor RoutingCongestion SuppressionPeer Mesh Coordination
Operates independently of cloud connectivity. Peer-to-peer node coordination only. |
| Slow Loop | Hours–Days | City-Wide OptimizationInfrastructure HealthSignal Plan UpdatesConfiguration Only
Cannot actuate signals directly. Cloud layer operates in advisory and configuration role exclusively. |
The architecture is not a hierarchy of override — it is a hierarchy of authority. The Fast Loop cannot be preempted by the Slow Loop. The cloud cannot command a signal directly. Each tier is structurally constrained to its domain by the physical architecture of the system, not by software policy that can be misconfigured, overridden, or compromised.
Who Commands the Signal Is Decided
by Physics, Not Policy.
The defining architectural property of UMIN is that control authority over physical signal hardware is determined by the physical wiring of the system — not by software rules, firmware configurations, or access control policies that can change.
The signal control interface — the hardware that governs what color a traffic light displays — is physically connected to exactly one command pathway: the direct hardware connection from the edge computing subsystem. There is no shared bus, no switch fabric, no network interface controller between any external communication system and the signal control interface. The cloud cannot command a signal. A network intrusion cannot command a signal. A firmware compromise in the communication module cannot command a signal. These are not access control policies. They are physical impossibilities given the wiring architecture.
This architecture has a second consequence for safety: if the edge computing subsystem fails to deliver a valid output within the hard latency bound, the signal controller transitions autonomously to a defined fail-safe state — implemented in hardware watchdog logic within the controller itself, independent of any processor, any firmware, and any network. All failure modes collapse into a single deterministic safety state. The system fails safe, not open.
“The intersection is not a passive piece of hardware. It is a governance boundary — and it must be governed at the speed of physics, not the speed of networks.”
Every Threat. Every Condition.
Every Second.
Situational awareness at the intersection requires sensing across multiple modalities simultaneously. No single sensor type is sufficient. UMIN deploys a multi-modal sensor array at each node, with all inputs fused in real time at the edge.
| Sensor | Primary Function | Safety Application |
|---|---|---|
| mmWave Radar | All-weather position and velocity tracking for vehicles and pedestrians | Collision trajectory prediction; red-light violation anticipation; pedestrian proximity sensing |
| RGB Camera | Object classification, lane occupancy analysis, and intersection geometry mapping | Vehicle type identification; queue density measurement; pedestrian detection in marked zones |
| Thermal Imaging | Biological entity detection under low-visibility: night, fog, heavy rain, glare | Pedestrian presence in complete darkness; wildlife crossing detection in rural deployment |
| Weather Array | Temperature, precipitation, road surface condition, atmospheric visibility | Ice and fog warnings to approaching vehicles; dynamic speed zone adjustments; hazard pre-positioning |
| Road Vibration | Structural anomaly detection, vehicle weight classification, surface degradation | Infrastructure health monitoring; overweight vehicle flagging; road condition alerts |
Sensor outputs are time-aligned and fused by the edge computing subsystem to produce a unified environmental state vector — a single consistent representation of the intersection environment at each inference cycle. This fusion architecture ensures that the AI model operates on synchronized, coherent scene data rather than asynchronous individual sensor streams.
Built to Prevent.
Not Merely to Record.
UMIN is not a surveillance system. It does not record identities. Its sole safety objective is preventing the next crash before it occurs — at the intersection, at speed, in real time.
The system predicts probable violations before they occur, using trajectory and velocity analysis. Signal timing can be extended or emergency alerts generated for approaching vehicles before a conflict forms.
Pedestrian presence triggers automatic signal timing extensions. Thermal imaging detects presence in conditions where cameras fail — nighttime, fog, rain. No pedestrian is invisible to the system.
Upon V2X authorization signal from an approaching emergency vehicle, the system pre-empts all current phase states and coordinates a clear corridor across multiple intersections simultaneously, with no manual dispatch required.
Adverse weather, ice, fog, and road surface anomalies detected at the node are transmitted to approaching connected vehicles via V2X. Speed guidance and hazard warnings reach drivers before the hazard is visible.
Thermal imaging detects large animals at crossing zones on rural highways, generating adaptive speed zone recommendations transmitted to approaching vehicles. An underserved safety category addressed at the infrastructure level.
Battery backup sustains full autonomous operation for 72 hours following grid power interruption. Life-safety functions — collision avoidance, signal control — continue without degradation during infrastructure outages.
Vehicles Become Participants,
Not Obstacles.
UMIN integrates with the Vehicle-to-Infrastructure (V2X) communication layer — the emerging standard for direct data exchange between road infrastructure and equipped vehicles.
Connected vehicles transmit real-time telemetry to UMIN nodes: speed, braking events, navigation intent, and hazard detection. The edge computing subsystem processes this data as First Loop input, incorporating vehicle behavior into the same real-time risk model that governs signal decisions. In the reverse direction, UMIN transmits Signal Phase and Timing (SPAT) data, map layer information, hazard warnings, and speed guidance recommendations to equipped vehicles.
UMIN supports both DSRC and C-V2X protocols and is architecturally positioned to integrate with autonomous vehicle systems as AV penetration of the vehicle fleet increases. The infrastructure does not require fully autonomous vehicles to deliver safety value today — it works with any V2X-equipped vehicle and continues to function for unequipped vehicles through infrastructure-level signal governance.
Intelligence Without Surveillance.
Public trust is a prerequisite for public infrastructure. UMIN is designed from first principles around anonymized, event-based data governance — not as a compliance layer, but as a core architectural constraint.
- ✓ Anonymous vehicle presence and trajectory events
- ✓ Aggregate flow density measurements
- ✓ Environmental and road condition data
- ✓ Anonymized event-type codes and timestamps
- ✓ Infrastructure health and sensor diagnostics
- ✗ License plate numbers or vehicle identifiers
- ✗ Driver identity or biometric data
- ✗ Individual behavioral or movement profiles
- ✗ Long-term individual travel history
- ✗ Facial recognition or camera-based identity matching
Individual identity resolution — if ever required for legitimate law enforcement purposes — is architecturally gated behind a conditional escalation protocol. It is not a default operating mode and is not accessible without cryptographically verified authorization. UMIN is an infrastructure intelligence system, not a surveillance platform.
Built to Scale from One Intersection
to a Region.
UMIN is designed for phased deployment that delivers measurable safety value at each phase, without requiring full-scale deployment to justify the investment.
Single-intersection Smart Pole Unit deployment. Basic edge AI signal optimization. Baseline safety and flow data collection. Proof-of-performance for city procurement.
Multi-intersection deployment along a defined corridor. Peer mesh coordination active. Emergency routing integration. Green wave optimization. First data monetization.
Full district deployment with V2X IoV integration. City Intelligence Cloud active. Safety contract activation. Third Temporal Control Layer city optimization live.
Highway and urban integration at regional scale. Full EV ecosystem coordination. Complete revenue stack active. Defense and sovereign infrastructure partnerships.
Each Smart Pole Unit is a modular, field-serviceable platform. Sensor, compute, and communication modules are replaceable without replacing the structural enclosure. The 20-year structural lifespan and 5-year technology upgrade cycle align UMIN with municipal procurement timelines rather than consumer electronics cycles.
A City That Sees Its Own Traffic
in Real Time.
The City Intelligence Cloud is the macro-intelligence layer of the UMIN system — a city-wide analytics and optimization platform fed by anonymized event data from the distributed node network.
The Traffic Prediction AI engine forecasts demand patterns and congestion formation before they occur. The Risk Heatmap Engine identifies geographic and temporal patterns in risk events. The Emergency Optimization Engine coordinates multi-intersection emergency corridors from a city-wide perspective, reducing emergency response times without requiring manual signal override from dispatch.
Critically, the City Intelligence Cloud holds no authority over individual signal states. Its outputs reach edge nodes exclusively through a configuration interface that is physically separated from the signal control interface — it updates behavioral thresholds and optimization parameters; it does not command signals. The governance architecture ensures that city-level intelligence enhances edge decision-making without creating the latency, single-point-of-failure, or attack surface risk that central control would introduce.
Quantified Outcomes.
Required for Procurement.
This section defines measurable performance targets for pilot deployment evaluation. All metrics are defined at intersection-node level and aggregated at corridor scale.
| Metric | Target | Category |
|---|---|---|
| Safety Performance | ||
| Total Intersection Collisions | 25% reduction (Year 1 pilot baseline) | Collision Reduction |
| Severe Injury Collisions | 35–50% reduction at high-conflict intersections | Collision Reduction |
| High-Probability Conflict Events | 40–60% reduction (AI-predicted trajectory) | Near-Miss Prevention |
| Emergency Response | ||
| EMS Travel Time | 20–35% reduction within pilot corridors | Dispatch-to-Arrival |
| Intersection Clearance | <8 seconds average for emergency corridor activation | Clearance Efficiency |
| Green-Wave Execution | 95% successful rate across connected intersections | Clearance Efficiency |
| Traffic Efficiency | ||
| Average Intersection Idle Time | 15–30% reduction | Flow Optimization |
| Corridor Throughput (Peak) | 10–25% improvement | Flow Optimization |
| System Performance | ||
| Edge Decision Latency | <100ms (hard requirement) | System Performance |
| Node Uptime (normal grid) | >99.5% | System Performance |
| Failover Activation Time | <300ms under subsystem degradation | System Performance |
Conservative modeling places ROI break-even at 3–7 years depending on corridor severity and baseline crash frequency. Annualized municipal savings include: reduced emergency response costs, reduced crash-related liability payouts, reduced congestion delay costs, and reduced insurance burden in pilot zones. See Section 20 for full ROI model.
Cost Framing Model.
Per-intersection deployment cost estimates for municipal capital budgeting, Phase 1 pilot configuration.
| Component | Estimated Cost Range |
|---|---|
| Smart Pole Retrofit Node (hardware + install) | $25,000 – $60,000 |
| Edge AI Compute Module | $5,000 – $12,000 |
| Sensor Suite (radar, thermal, optical) | $8,000 – $20,000 |
| V2X Communication System | $3,000 – $10,000 |
| Integration + Calibration | $10,000 – $25,000 |
| Total Estimated Cost per Intersection | $50,000 – $125,000 |
Pilot Corridor Example — 10 Intersections
Failure Mode and
Safety Degradation Architecture.
The system is designed so that no failure mode results in unsafe signal behavior. At no point does system failure produce uncontrolled or non-deterministic signal states.
| Level | Failure Mode | System Response |
|---|---|---|
| L1 | Edge AI Failure | System reverts to deterministic fixed-cycle traffic timing. No predictive control is executed. |
| L2 | Sensor Degradation | Remaining sensors continue operation. System reduces confidence weighting and defaults to conservative timing mode. |
| L3 | Full Edge Compute Failure | Signal controller enters pre-defined hardware-safe cycle mode. Independent watchdog circuit enforces timing fallback. |
| L4 | Communication Loss | System continues fully autonomous local operation. No dependency on external network exists for safety function. |
| L5 | Power Loss | 72-hour battery backup sustains signal control, basic sensing, and emergency corridor functionality. |
At no point does system failure result in uncontrolled or non-deterministic signal behavior. All degradation paths terminate in a defined, hardware-enforced safe state.
Procurement Language Reference.
Compliant terminology for DOT submissions, RFP responses, and government procurement review. Replace consumer-facing AI language with the DOT-standard framing below.
| Replace (Avoid) | Use Instead (Compliant) |
|---|---|
| AI-controlled intersections | Edge-computed adaptive traffic control infrastructure with deterministic safety constraints |
| Real-time predictive system | Infrastructure-based real-time situational awareness and hazard mitigation system |
| Surveillance data processing | Event-based anonymized traffic state estimation system |
| Autonomous decision-making system | Locally governed signal optimization system with hardware-enforced safety constraints |
| City-wide AI traffic network | Distributed transportation infrastructure coordination network |
Procurement Positioning Statement.
UMIN is a distributed transportation infrastructure upgrade system designed to improve intersection safety, reduce emergency response latency, and optimize traffic flow through edge-computed, event-based intelligence systems. The system is explicitly designed to operate within existing municipal governance structures, with strict separation between safety-critical signal control functions and higher-level analytical and optimization systems. UMIN introduces no dependency on continuous cloud connectivity for life-safety operations and enforces deterministic fail-safe behavior at the hardware level. All identity-related data processing is excluded from routine operation and is only accessible under lawful, warrant-based authorization procedures.
Competitive Landscape and
System Displacement Analysis.
UMIN is not an incremental optimization layer for existing adaptive signal systems. It is a distributed edge-control architecture designed for sub-second safety response at the intersection level.
Comparative System Assessment
| System Class | Primary Design Objective | Documented Constraint | UMIN Architectural Difference |
|---|---|---|---|
| SCOOT / SCATS (Central Adaptive) | Network-level traffic optimization | Centralized processing with multi-intersection coordination and update intervals measured in minutes | UMIN performs intersection-local computation with sub-100ms response at the edge node, independent of centralized scheduling loops |
| Camera-Based Smart Signal Systems | Visual traffic density estimation | Performance degradation under low visibility conditions (fog, glare, occlusion) | UMIN integrates multi-modal sensing (radar, thermal, optical) to reduce single-sensor dependency |
| Cloud-Based AI Traffic Management | Fleet-wide optimization via centralized AI models | Network latency and dependency on external connectivity for decision propagation | UMIN safety-critical control logic operates locally at the intersection edge node without reliance on cloud round-trip communication |
| Fixed-Time Signal Controllers | Predefined timing cycles | No real-time adaptation to dynamic traffic or emergency conditions | UMIN replaces static timing dependency with real-time sensor-driven state evaluation |
| V2X-Dependent Coordination Systems | Vehicle-to-infrastructure cooperative control | Dependent on vehicle penetration rates and onboard system availability | UMIN provides infrastructure-side control logic applicable to all vehicles, independent of V2X adoption |
Retrofit Integration Model
| Infrastructure Element | Integration Method | Deployment Impact |
|---|---|---|
| Signal Controllers (NEMA TS-2 / ATC) | Interface via standard controller communication protocols | No replacement required |
| Existing Signal Heads and Poles | Maintained; actuation commands passed through existing outputs | Physical infrastructure retained |
| Power Systems | Standard municipal electrical supply with optional backup support | No baseline electrical redesign required |
| Traffic Management Centers (TMCs) | Data ingestion via standard traffic protocol interfaces (NTCIP-compatible feeds) | Coexistence with existing systems |
| Emergency Management Systems | Optional integration via structured data interface layer | Additive integration only |
UMIN represents a new control intelligence layer rather than a replacement of physical assets. It is designed to operate alongside existing signal investments, preserving prior capital expenditure while adding edge-distributed safety intelligence.
Distributed Control Authority and
Remote Access Elimination.
This section defines the control authority model governing UMIN operation, with emphasis on elimination of externally routable actuation pathways for safety-critical signal control.
Control Pathway Model
Traffic analytics. Non-safety optimization inputs. Advisory-level data exchange only. No actuation authority.
Data ingestion only. No actuation authority. No safety-critical command pathway access.
Signal Actuation Flow
↓
[Edge Fusion Processing — Layer 1]
↓
[Deterministic Safety Decision Engine]
↓
[Local Signal Actuation Interface]
↓
[Physical Signal State Change — Layer 0]
——————————————————————————————
External systems: READ-ONLY DATA FLOW ONLY
No reverse path: Layer 3 → Layer 0 does not exist
Security and Access Model
| Component | Exposure Type | Access Level |
|---|---|---|
| Signal Actuation Hardware | Physically isolated | LOCAL ONLY |
| Edge Control Node | Local network only | RESTRICTED |
| Municipal Systems | Network-accessible | READ-ONLY |
| Cloud Systems | Internet-accessible | OBSERVATIONAL |
Verification, Validation, and
Certification Pathway.
This section defines the verification methodology, validation requirements, and certification pathway for deployment of UMIN systems in municipal traffic infrastructure environments.
Three-Method Verification Framework
- Inspection of hardware wiring architecture
- Confirmation of control pathway isolation
- Validation of fail-safe relay behavior
- Live intersection behavior under controlled traffic conditions
- Measured response latency under real-world vehicle/pedestrian loads
- Emergency pre-emption performance validation
- Digital twin modeling of intersection behavior
- Stress testing under high-density and failure conditions
- Scenario-based validation of control loop stability
Performance Acceptance Thresholds
| Category | Minimum Acceptance Threshold |
|---|---|
| Collision Reduction | Statistically significant reduction vs. baseline intersection data |
| Emergency Response Improvement | Measurable reduction in corridor traversal time |
| System Stability | >99% operational continuity during normal grid conditions |
| Fail-Safe Activation | 100% deterministic safe-state transition under induced fault conditions |
| Control Isolation | Verified absence of external actuation pathways |
Certification Outcome States
All thresholds met or exceeded. No unresolved safety exceptions.
Partial threshold achievement. Additional data required for scaling decision.
Failure in control isolation, latency, or fail-safe behavior.
Chicago Corridor Model —
Pilot Deployment Framework.
This section defines the structure, scope, and evaluation criteria for a controlled pilot deployment of the UMIN system within a selected urban corridor in Chicago, Illinois. The pilot is a performance validation corridor, not a production deployment.
Proposed Pilot Geography
- Stony Island Ave & 79th St
- Western Ave & Madison St
- Cicero Ave & Chicago Ave
- Ashland Ave & 63rd St
- Michigan Ave & Roosevelt Rd
- Lake Shore Dr & Belmont Ave
- State St & Congress Pkwy
- Halsted St & North Ave
- Irving Park Rd & Pulaski Rd
- 95th St & Dan Ryan Expressway ramps
Deployment Phases and Timeline
Electrical and controller compatibility validation. Signal timing baseline capture. Traffic volume and conflict zone mapping. Sensor placement modeling. Deliverable: Intersection digital twin models + baseline performance metrics.
Installation of edge nodes and sensor arrays. Integration with existing signal controllers. Calibration. Safety state verification testing. Deliverable: Fully instrumented pilot intersections + operational readiness report.
Real-time operation under municipal traffic conditions. Continuous performance logging. Emergency vehicle routing tests. Side-by-side baseline comparison. Deliverable: Live performance dataset + safety and latency validation report.
Independent audit review. Municipal performance assessment. Safety and reliability scoring. Expansion recommendation report. Deliverable: Pilot outcome certification package.
Risk Management Framework
| Risk Category | Mitigation Strategy |
|---|---|
| Hardware failure | Redundant fail-safe relay architecture |
| Sensor degradation | Multi-modal sensor fusion redundancy |
| Traffic disruption during deployment | Staged rollout per intersection |
| System incompatibility | Pre-installation compatibility audit |
| Operational uncertainty | Parallel baseline comparison system |
Total Cost of Ownership, ROI Modeling,
and Federal Funding Alignment.
This section translates quantified performance outcomes and deployment costs into a unified financial model suitable for municipal capital budgeting, state DOT infrastructure planning, and federal transportation safety grant evaluation.
Pilot Corridor Economic Model — Chicago 10-Intersection Reference
| Category | Annual Value Impact (Conservative) |
|---|---|
| Collision reduction savings | $2M – $8M |
| Emergency response efficiency | $0.5M – $2M |
| Traffic efficiency gains | $1M – $5M |
| Liability and insurance reduction | $1M – $6M |
Federal Funding Program Alignment
| Program | Agency | Alignment |
|---|---|---|
| Safe Streets and Roads for All (SS4A) | USDOT | Intersection safety modernization; pedestrian and vulnerable road user protection |
| SMART Grants | USDOT | Intelligent transportation systems deployment; data-driven infrastructure improvement |
| Highway Safety Improvement Program (HSIP) | FHWA | High-crash corridor remediation; measurable safety outcome requirement |
| IIJA ITS Deployment | Congress / FHWA | Connected and automated infrastructure; V2X integration programs |
| CMAQ Program | FHWA / FTA | Congestion mitigation; air quality improvement through traffic flow optimization |
System Architecture Diagrams and
Control Boundary Definition.
This section defines the physical and logical structure of the UMIN system through standardized engineering architecture descriptions for DOT engineering review packages, ITS architecture submissions, and pilot procurement documentation.
System Architecture Overview — Logical Layer Model
Traffic signal heads (red/yellow/green). Electrical relays and actuation wiring. Pole-mounted infrastructure hardware. Executes final physical signal state.
Traffic Management Center (TMC). Corridor analytics systems. Historical data processing. Observational and advisory only — no actuation authority.
Cloud analytics platforms. Regional coordination systems. Third-party mobility data. Data ingestion and reporting only.
Failure State Architecture — Logical Behavior Model
| Scenario | Failure Condition | System Behavior |
|---|---|---|
| A | Sensor Failure | System reverts to multi-sensor redundancy weighting. Signal state maintained via deterministic fallback logic. |
| B | Edge Node Degradation | System transitions to predefined safe timing mode. Physical signal relay maintains operational continuity. |
| C | Network Disconnection | No operational impact on signal actuation. System continues independent local operation indefinitely. |
| D | Power Instability | Hardware relay enters fail-safe timing state. Signal behavior transitions to controlled deterministic cycle. |
Emergency Priority Control Pathway
↓
[Edge Priority Override Logic — Layer 1]
↓
[Intersection Clearance Sequencing]
↓
[Corridor Synchronization Broadcast — Peer Mesh]
↓
[Signal State Coordination Across Adjacent Nodes]
// This pathway is local-first and does not depend on external network confirmation
The defining architectural constraint of UMIN is: Safety-critical signal actuation is physically bound to local edge control hardware and is not dependent on external network systems, cloud platforms, or centralized command infrastructure. This boundary is enforced through physical wiring topology, control logic segmentation, and execution-level isolation.
The Infrastructure of Safety
Is Long Overdue.
UMIN represents a fundamental change in the relationship between road infrastructure and the lives it governs. Roads have been passive for a century. The technology to make them active, aware, and protective exists today. The only remaining question is the speed of deployment. Maven Prodigy Algorithmic Corporation invites discussion with municipal governments, transportation agencies, infrastructure investors, and technology partners who share the conviction that the cost of inaction is measured in lives.