CUMRA: Smart Parking
Discover how CUMRA teamed up with CEMET to investigate whether outdoor parking spaces could be identified and measured using cutting-edge mapping data, vehicle information APIs, and minimal on-street hardware
CUMRA is an ambitious innovation company dedicated to tackling real-world urban challenges through emerging technologies. With a background in digital product development, Cumra approached CEMET with a compelling concept: to explore how intelligent systems could be used to identify and monitor outdoor parking spaces more effectively — addressing a persistent issue faced by towns and cities across the UK.
Locating and managing parking in busy urban environments continues to be a daily source of frustration for both drivers and local authorities. Cumra wanted to explore whether a scalable, low-cost solution could be developed to automatically detect on-street and off-street parking spaces — determining not only whether a space is available, but also if it’s suitable for a particular type of vehicle. The challenge was to assess the feasibility of such a system using existing data sources and minimal hardware, reducing reliance on expensive infrastructure from the outset.
What Did We Do?
CEMET delivered a six-week research and development project structured across three focused sprints. This work explored the technical landscape, identifying both the opportunities and limitations involved in creating an intelligent parking detection system.
Our research examined a range of off-the-shelf number plate recognition APIs, exploring their suitability for early-stage development. We also investigated publicly available vehicle data through the DVLA, alongside richer commercial alternatives that could provide key dimensions such as vehicle length, width and height. These data points were critical in assessing how technology could determine if a parking space is suitable for a specific vehicle.
We evaluated the use of geospatial services including HERE Technologies and Google Street View to understand how live mapping data might be used to estimate parking availability, and considered whether this could be done without installing physical sensors. At the same time, we assessed hardware options such as the Jetson Nano and Raspberry Pi, identifying the Jetson Nano as the more capable platform for computer vision-based tasks.
To better understand environmental constraints, we explored the potential of LIDAR and depth camera systems for identifying empty spaces and estimating parking bay sizes, as well as testing the feasibility of using live webcam footage to cross-reference map-based data. All findings were compiled into a comprehensive Jupyter Notebook Knowledge Transfer Document, providing Cumra with a detailed set of next steps.
By the end of the feasibility study, CUMRA had a clear, evidence-based roadmap for progressing towards a working proof of concept. Our guidance included practical recommendations on selecting number plate recognition and vehicle data APIs, advice on suitable hardware platforms, and a deep evaluation of how mapping technologies could be leveraged before any physical installation takes place. The project also highlighted how CUMRA might begin by integrating existing live camera feeds to test the accuracy of map data, while exploring more advanced sensor options like LIDAR and depth cameras for future scaling. With this foundation in place, CUMRA is now well-positioned to take the next step in building an intelligent, scalable parking solution — one that could transform the way we navigate and manage public spaces.
With a solid research foundation now in place, Cumra are ready to explore next-stage development and move one step closer to smarter, more efficient parking solutions. To find out more about their latest work visit their website.