Innovative tool for predicting and controlling through Deep Learning
TOMMs-AI leads an emerging transport solution
- TOMMs-AI adopts the artificial intelligence to predict what is going on in the near future and provides proactive solutions.
- CNN(Convolutional neural network) family creates feature maps through the input of hidden layers. Therefore, many traffic-related data (primarily spatial and temporal data) are carefully prepared for the convolutional kernels, and different model depth was introduced to match corresponding feature learning needs.
- The sequential computation of RNN(Recurrent neural network) family, is proposed for better temporal features learning. Furthermore, the structural joining of CNN and RNN appears to be more effective for learning both spatial and temporal features.
- TOMMS-AI contains multiple components for specific learning purpose. The developed solution has the consideration of both latent features and the working mechanism of neural networks.
At a glance
- Proactive management
- Efficient control in demand and signal plans
- Advanced prediction tool
State of the Art
Vast research has already proved that neural network-based solutions outperform traditional machine learning solutions
Prediction
A combination of CNN and RNN allows extracting spatial and temporal features more effectively so that increase the accuracy and decrease the error in the model
Control
Reinforcement learning helps the control to be stable. The Google Deep Mind team proves that the deep learning is the next generation approach by Alpha Go’s win
Fast solution
The cloud service ensure a fast solution and secure all the data collected
Intelligent approach
If we predict what is going on in near future, we are able to control the traffic in a smart way. The TOMMs-AI makes this happen
The advanced prediction models inspired by AI
