AI Assisting in Traffic Relief


Due to the increasing number of cars, frustration accompanying drivers searching for free parking spot is escalating. During our tutorial we would like to present a practical approach to implement a real-time smart vision parking solution targeted to improve parking experience. The solution is based on applying object detection model on real-time video stream captured by city cameras and providing information on the number of available parking spots and their location. Thus the system allows for a significant reduction in time wasted by drivers on wandering around, which generates extra traffic. We lead the audience through the multiple-stage process of solution deployment assuming some level of latency required by this type of application. The first stage encompasses a GPU-based training of a state-of-the-art object detection model conducted both on SAS Viya platform using DLPy package and Pytorch package within SAS Open Analytics Platform embracing wide range of open source technologies into its ecosystem. During the following step, we show how to design an end-to-end pipeline of stream processing logic in SAS Event Stream Processing with a particular focus on deploying a deep learning model trained in the earlier step. Finally, we present how easy it is to embed a streaming data to generate dashboard for monitoring demand dynamics and traffic level, which might be extremely useful for a traffic control department. However, the key end user is the driver, therefore, as the last stage we show a mockup of a mobile app which sends requests for information on parking spots availability to a streaming engine whilst the driver is approaching a parking area and uses the system’s response in guidance.

Background Knowledge
basic knowledge of computer vision and machine learning is required, no SAS tools knowledge is required


Piotr Kaczyński is a Senior Business Solutions Manager in the analytical team. Responsible for business development in the area of ​​analytics and system integration, also at the interface between SAS and open source tools. For almost 8 years, as a researcher, he conducted research on the convergence of linearization algorithms for stochastic processes. Since 2003, he has been associated with analytics through the practical use of predictive modeling for forecasting in the energy sector. He also participated in projects in the field of data mining, BI and forecasting for sea transport. His main interests are artificial intelligence, and in particular its practical use in image processing.

Open Data Science




Open Data Science
One Broadway
Cambridge, MA 02142

Privacy Settings
We use cookies to enhance your experience while using our website. If you are using our Services via a browser you can restrict, block or remove cookies through your web browser settings. We also use content and scripts from third parties that may use tracking technologies. You can selectively provide your consent below to allow such third party embeds. For complete information about the cookies we use, data we collect and how we process them, please check our Privacy Policy
Consent to display content from - Youtube
Consent to display content from - Vimeo
Google Maps
Consent to display content from - Google