Reducing road congestion through intelligent Urban Traffic Control (UTC) 
is desirable due to the economic, environmental and social costs of 
congestions. An efficient solution to UTC must be adaptive in order to 
deal with the highly-dynamic nature of urban traffic. In the near 
future, increased instrumentation such as global positioning systems and 
vehicle-to-vehicle/infrastructure communication will provide us with 
detailed traffic information. We describe  a next-generation UTC system 
where junction controllers exploit this knowledge about traffic at 
individual junctions to optimize signal control. In addition, 
neighbouring junctions collaborate with each other to improve global 
performance. Finally, we present a non-parametric pattern-change 
detection mechanism that triggers relearning when traffic pattern 
changes significantly.