ITMS
Intelligent Traffic Management System
ITMS applies reinforcement learning, atmospheric intelligence, and multi-modal sensing to urban traffic management. It moves signal control from fixed-time plans toward adaptive, continuously learning systems that respond to real-world conditions — improving flow, safety, and equity across the full spectrum of urban movement.
Urban traffic is not a fixed problem. It is a dynamic system that demands learning intelligence.
Traffic signal control is one of the oldest optimization problems in applied engineering — and one of the least transformed by modern AI. Most urban intersections still operate on timing plans configured by hand, calibrated annually, and unable to respond to the second-by-second variation in demand that determines actual traffic performance. The gap between fixed-plan control and optimal adaptive control represents billions of person-hours of unnecessary delay annually.
ITMS is Jaqlor's research effort to close this gap. We develop adaptive signal control systems grounded in reinforcement learning and atmospheric intelligence, test them in high-fidelity simulation environments, and build toward real-world validation in partnership with urban infrastructure operators. The goal is not marginal improvement — it is the elimination of fixed-plan traffic management as the default paradigm.
Traffic congestion is not an inevitable feature of urban life — it is a systems failure, sustained by infrastructure that operates on timing plans written years ago for conditions that no longer exist. Every minute of unnecessary delay in emergency response carries a measurable human cost. ITMS is being built on the belief that cities deserve traffic systems as intelligent and adaptive as the people who move through them.
Six-layer adaptive traffic intelligence stack
Traffic Sensing & Detection Layer
Multi-modal vehicle detection from inductive loop sensors, radar, LiDAR, and video analytic feeds. Pedestrian and cyclist detection via computer vision pipelines. Real-time occupancy, speed, and queue length estimation at each monitored intersection and road segment.
Network State Estimation
Fusion of sensor streams, probe vehicle GPS data, and historical flow models generates real-time network-wide traffic state estimates. Kalman filter-based state estimation handles sensor drop-outs and gaps in coverage. Incident detection via statistical process control on flow time series.
Atmospheric & Environmental Integration
AtmosPy atmospheric intelligence feeds incorporated for visibility estimation, road surface condition inference, and adverse weather routing. Environmental context modulates signal timing plans, advisory speeds, and incident response protocols.
Signal Optimization Engine
Reinforcement learning agents trained on traffic micro-simulation environments optimize signal timing plans for throughput, delay, and emissions objectives. Multi-objective optimization handles competing priorities between arterial throughput and pedestrian crossing safety.
Adaptive Routing & Coordination
Network-level coordination protocols distribute traffic across parallel routes to prevent localized oversaturation. Cooperative signal green wave propagation on high-demand corridors. Emergency vehicle preemption with automated arterial clearance.
Intelligence Delivery & Integration
Structured APIs deliver traffic state, predictions, and recommendations to navigation platforms, transit operators, and emergency services. NTCIP-compatible protocols for existing traffic management center integration. Real-time performance dashboards and operator interfaces.
What ITMS is actively investigating
Reinforcement Learning for Adaptive Signal Control
Traffic signal control is a sequential decision problem in a non-stationary environment — every cycle creates a new state, and decisions interact across intersections through vehicle queuing dynamics. Classical fixed-time and actuated control plans are calibrated to historical demand patterns and fail gracefully when conditions deviate. The transition from rule-based to learning-based control represents the most significant opportunity for urban traffic efficiency improvement in a generation.
ITMS research develops multi-agent reinforcement learning frameworks where each intersection controller learns adaptive timing policies through interaction with simulation environments calibrated to real-world traffic data. We study coordination protocols that allow neighboring intersections to share state and synchronize decisions, enabling green wave propagation and corridor-level optimization without centralized control architecture.
Weather-Adaptive Traffic Management
Precipitation, fog, ice, and extreme heat are among the most significant disruptors of urban traffic flow — and among the least well-handled by conventional management systems. Standard signal timing plans and advisory speed limits are static, designed for dry, clear conditions. The result is systematic underperformance precisely when safe, efficient movement matters most.
Through integration with AtmosPy, ITMS develops weather-aware traffic management that responds dynamically to atmospheric state. Our research models how visibility impairment, wet surface friction, and reduced driver attentiveness affect speed-flow relationships and safe headway requirements. Signal timing plans, advisory speeds, and routing recommendations adapt in real time to atmospheric conditions, improving safety outcomes and maintaining throughput during adverse weather events.
Equitable Traffic System Design
Algorithmic traffic optimization has historically treated vehicles as homogeneous agents and throughput as the primary objective. This framing systematically underweights pedestrian and cyclist movement, disadvantages neighborhoods with lower vehicle ownership, and concentrates optimization benefits in high-traffic corridors at the expense of residential streets that bear induced traffic.
ITMS research develops multi-modal, multi-objective optimization frameworks that explicitly represent pedestrians, cyclists, transit vehicles, and emergency responders alongside private vehicles. We study how optimization objective weighting affects distributional outcomes across different urban neighborhoods and demographic groups, aiming to build adaptive systems that improve mobility equity rather than simply maximize aggregate throughput.
Cooperative Infrastructure-to-Vehicle Coordination
The emergence of connected and semi-autonomous vehicles creates an opportunity for infrastructure and vehicle intelligence to cooperate in ways impossible with passive detection systems. Infrastructure-to-vehicle communication allows traffic management systems to provide real-time signal phase and timing data, routing guidance, and hazard warnings directly to vehicle control systems.
ITMS research investigates the cooperative control protocols that enable this coordination — how signal controllers and vehicle navigation systems can jointly optimize corridor performance in mixed traffic environments containing connected, partially autonomous, and conventional vehicles. We study how penetration rate affects system-level benefit and how protocols must adapt to remain effective at low connected vehicle market shares.
From sensor signal to optimized traffic flow
Intelligence Pipeline
Multi-Modal Traffic Detection
Loops, radar, LiDAR, and video feeds capture occupancy, speed, and queue state in real time.
Network State Fusion
Sensor streams fused with probe vehicle data to generate network-wide traffic state estimates.
Environmental Context Integration
AtmosPy atmospheric intelligence incorporated for weather-adaptive signal and routing decisions.
RL Signal Optimization
Multi-agent reinforcement learning computes adaptive timing plans for intersections across the network.
Coordination & Routing
Network-level green wave propagation and route advisory generation for congestion distribution.
Intelligence Delivery
Optimized signal plans, operator dashboards, and navigation platform data distributed to downstream systems.
Where ITMS intelligence is deployed
Simulation-Based RL Training
High-fidelity SUMO traffic simulation environments parameterized with real-world network geometries and demand profiles. Multi-agent RL policies trained and benchmarked against conventional actuated and fixed-time control baselines.
AtmosPy Environmental Layer
Integration with AtmosPy atmospheric intelligence for weather-adaptive management. Precipitation, visibility, and road surface condition data incorporated into signal timing and routing recommendation logic.
Corridor Performance Analytics
Automated performance measurement pipeline for arterial corridors. Probe vehicle data processed into travel time, delay, and stop rate metrics. Anomaly detection identifies performance degradation events for targeted intervention.
Emergency Vehicle Preemption
Integrated emergency vehicle detection and preemption system. GPS-based EV tracking activates preemption sequences upstream of approaching vehicles; automated corridor clearance minimizes response time while managing network-level disruption.
Path to autonomous urban traffic coordination
Signal Intelligence
RL adaptive signal control, simulation environments
Adaptive Routing
Network coordination, weather integration, multimodal
Optimization
City-scale deployment, equity objectives, emissions
Autonomous Coordination
V2I cooperative control, connected vehicle protocols
Intelligence Domain
Environmental & Urban Systems→Related System
AtmosPy→Research partnerships & collaboration
Jaqlor engages with research institutions, industry partners, and technical organizations seeking rigorous collaboration in urban mobility and traffic intelligence research. Partnerships prioritize technical depth, defined scope, and responsible development.