Pictured from L-R: Rodrigo Ordonez-Hurtado (UCD), Giovanni Russo (IBM), Wynita Griggs (UCD), Martin Mevissen (IBM), Julien Monteil (IBM), Jonathan Epperlein (IBM), Yassine Lassoued (IBM) and in the car Prof. Robert Shorten (UCD)
The Cognitive In-Car Companion is designed to look after the driver and the passengers while being as unobtrusive as possible. Using historical data as well as real-time data from in- and outside the car, the companion predicts the driver’s intentions while being aware of the environment and the context of the driver’s journey. If the companion foresees any difficult or risky situations, it assists the driver in avoiding them.
Imagine you are getting a late start to work at the UCD campus. Due to local parking regulations, your usual parking lot is not accessible from your usual entrance onto campus anymore; because you usually arrive to work much earlier than today, you are not aware of these regulations — but the companion is.
Having predicted your most likely destination and time of arrival and realizing the relevance of the regulations, the companion will speak up and suggest that you enter campus through an alternative gate .
To fulfil its duty to provide only relevant information while not bothering the driver otherwise, the Cognitive In-Car Companion draws upon a parsing component, which filters the available streams of data to extract the data that is relevant in the context of the current journey. The context is provided by a route prediction component, which is always trying to predict the driver’s intentions based on the output of a Markov model trained and updated based on the history of trips taken by the driver. Finally, a risk assessment component ingests the parsed data streams to detect potential risks along the journey. If – and only if – the confidence in its prediction of the route and an associated risk are high enough, the companion springs to life and assists the driver.
In this way, the driver never has to concern themselves with any of the data the companion takes into consideration to provide assistance, letting the companion do the necessary work.
While the prediction, parsing, and risk assessment are based on algorithms developed by the researchers at IBM and UCD, the interaction between companion and driver leverages cloud services available through IBM Bluemix: it uses Watson Speech to Text to listen to the driver, Watson Conversation’s natural language processing to understand him or her, and Watson Text to Speech to respond.
A key motivation for this project is the increasing amount of data available inside a car, coming from a wide range of data sources. This could be environmental data, data about the current state of the car, data exchanged by other cars, or data about preferences and past behaviour of the driver and passengers.
The challenge – which is also facing automotive manufacturers – is to be able to identify, within these massive streams of data, the pieces of data that are relevant to the driver in the context of the current journey. The companion from IBM Research – Ireland in collaboration with UCD, sifts through the data streams, filters out irrelevant data, and only provides relevant information to the user.
Such parsing engines have great potential for other industries facing the challenge of data overload as so many are. Cognitive systems like the companion can augment human abilities by unlocking insights from data that would otherwise remain hidden.
This project has received funding from the Electronic Component Systems for European Leadership Joint
Undertaking under grant agreement No 692455. This Joint Undertaking receives support from the European
Union’s Horizon 2020 research and innovation programme and Austria, Denmark, Germany, Finland, Czech
Republic, Italy, Spain, Portugal, Poland, Ireland, Belgium, France, Netherlands, United Kingdom, Slovakia,
Norway. The Irish use case of this project is also co-funded by Enterprise Ireland and IDA.
It is no surprise that following the massive success of deep learning technology in solving complicated tasks, there is a growing demand for automated deep learning. Even though deep learning is a highly effective technology, there is a tremendous amount of human effort that goes into designing a deep learning algorithm.
For the second year in a row, researchers at the IBM-Illinois Center for Cognitive Computing Systems Research(C3SR) won a competition challenging experts worldwide to design low-power embedded systems for Internet-of-Things (IoT) applications. The 2019 Design Automation Conference (DAC) System Design Contest’s objective: create algorithms that can accurately detect and locate objects from images taken by […]