WISP provides an AI-driven platform that uses deep reinforcement learning to continuously optimize traffic light sequences across city intersections, reducing vehicle idle time and emissions. The cloud‑based solution works with existing signal infrastructure, offering real‑time adaptation, travel‑time analytics, and predictive wait‑time estimates for municipal traffic agencies.
Funding
Funding not disclosed
Founders
Product
Problem
Urban traffic congestion leads to long vehicle idle times, increased emissions, and reduced air quality in large cities. Existing traffic signal systems are static and cannot adapt in real time to fluctuating traffic patterns, limiting overall efficiency.
Solution
WISP offers an AI-driven platform that dynamically optimizes traffic light sequences across multiple intersections using a deep reinforcement learning agent. The system continuously learns the complex relationship between network congestion and signal phasing, adjusting timings to minimize stop‑and‑go conditions. By operating on existing traffic infrastructure, it reduces vehicle idle time, cuts CO₂ emissions, and improves air quality without the need for additional hardware. Deployed in several municipalities, the platform provides measurable travel‑time savings and pollution reductions, and it can predict remaining wait times for drivers stopped at red lights.
Target Audience
Primary customers are municipal traffic management agencies and city transportation departments seeking to improve traffic efficiency and environmental outcomes in dense urban areas.
Features
- Deep Q‑Learning reinforcement learning agent that selects optimal signal phase sequences for coordinated intersections
- Real‑time adaptation to traffic flow changes without requiring new roadside sensors or equipment
- Cloud‑based analytics that quantify travel‑time savings, CO₂ reduction, and noise‑pollution impact
- Predictive module estimating remaining wait time for each stopped vehicle
- Scalable architecture capable of being applied to any urban road network