Performance of pheromone model for predicting traffic congestion
AAMAS '06 Proceedings of the fifth international joint conference on Autonomous agents and multiagent systems
Generalized Additive Models (Texts in Statistical Science)
Generalized Additive Models (Texts in Statistical Science)
SPADE: the system s declarative stream processing engine
Proceedings of the 2008 ACM SIGMOD international conference on Management of data
Driving with knowledge from the physical world
Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining
Discovering spatio-temporal causal interactions in traffic data streams
Proceedings of the 17th ACM SIGKDD international conference on Knowledge discovery and data mining
Online Learning Solutions for Freeway Travel Time Prediction
IEEE Transactions on Intelligent Transportation Systems
Hi-index | 0.00 |
Transport authorities have been deploying and utilising sensor infrastructures in order to improve upon the level of transport-related services within cities. As existing resources are more and more constrained, novel means of utilising the data originating from these sensors are sought. Advanced IT infrastructures enable real-time processing use cases, such as travel time estimations along defined corridors. One major challenge is to deploy general purpose machine learning algorithms that are able to learn relationships between the covariates and a defined response variable. We introduce a data fusion approach using generalised additive models (GAM) to estimate journey times online in a real-time streaming platform. We experiments with bluetooth sensors and weather information to improve the estimation of journey times along a defined A3 corridor in south-central London. Our approach is able to continuously improve the journey time estimation as new (high-frequency) data becomes available. Our fusion platform also generalises to be able to process more data sources and it scales to city-wide deployments. This way, existing legacy sensor deployments can be utilised for novel value-added services and investments into infrastructural sensor deployments can be assessed in a data-driven way and on a needs basis.