May 1, 2023 · Campus
Capstone Project 2022/23 | AI for the Car Wash Facility
The Digital Business & AI Management major within the Bachelor of Business Administration at HWZ concludes with a capstone project. In this project, students are required to develop a business case for the use of AI within a company. This year, the best project will be recognised with an award. You can find out more about the winning project and the award in the interview with the winning team.
About the award
In 2019, the Digital Business & AI Management major within the Bachelor of Business Administration was developed and offered for the first time in autumn 2020. For the past three years, bachelor’s students at HWZ have gained in-depth insights into selected aspects of artificial intelligence, whose influence and impact continue to grow. The award was established in honour of the most influential co-founder of this specialisation, who sadly passed away at the beginning of 2022. This award recognises qualities that reflect the co-founder’s character: innovative strength, sustainability, and social value.
The winning project
An AI for the car wash conveyor system
In a conveyor car wash, a vehicle can be processed every 30 seconds. In contrast, in a gantry car wash, vehicles remain stationary during the washing process. On average, this washing process takes 10.5 minutes.
Problem
At peak times (for example, on days with Saharan dust, public holidays, or fine Saturdays), portal car washes often experience very long waiting times. This is because the washing time is fixed, regardless of how dirty the vehicles are. Whether the vehicles are only lightly or heavily soiled, the washing time is 10.5 minutes. As a result of the waiting times, customers leave the car wash.
Objectives
The primary objective of a gantry car wash is to ensure customer satisfaction at all times. Future improvement measures should be designed to continue delivering flawless washing results. Operators should be able to fully realise their revenue potential.
Solution
Depending on the type of contamination, the rotating washing brushes require varying amounts of time to remove the dirt. A computer vision AI model is to be used to optimise the processing time of vehicles in automatic car washes. This is achieved by having the AI model predict the likelihood of different types of contamination based on image data. A subsequent set of rules then regulates the processing time according to the model’s output via software programming.
Annina Berweger, Daniel Eng and Lukas Huber were responsible for the winning group. We met them for an interview to learn more about the project from their perspective.
What is your connection to car wash facilities? How did you become aware of the problem?
We looked for a practical example from our professional experience. A team member owns a petrol station and car wash in the heart of Zurich Wiedikon and spoke about challenges with automatic car wash systems. This led us to the issue that the washing time is always the same and does not take into account the actual level of dirt on the vehicle. As a result, his car wash also experienced queues and a loss of revenue on fine summer days, as customers left due to excessive waiting times.
How did you approach finding a solution?
It was an iterative process. We researched best practices in the field of computer vision, gathered possible solution approaches, and examined which hardware (camera/lighting) could be used. The exchange with the lecturers was very helpful in this regard.
What was the greatest challenge in the project?
Our initial solutions were very comprehensive. They also included the control of the washing nozzles and the addition of chemicals depending on the level of contamination, and were correspondingly complex. It was important to reduce this complexity and focus initially on the essentials. As a result, detecting the degree of contamination on the vehicle became the main priority. All other features were left in the backlog.
To what extent did the expertise gained from your studies so far help you with this project?
In the first stage, we assessed the economic viability of the project, and in the second stage, we focused on a potential technological solution. The knowledge gained from our main degree programme in business economics and the major in digital business and AI management was crucial in this process.
What would you say is your most significant takeaway?
An interdisciplinary project team is essential. In addition to entrepreneurial thinking, expertise from various fields—such as data science, software engineering, project management, photography, and lighting—is required to effectively implement such a project. Furthermore, a complex problem must be simplified in order to bring an MVP to market as quickly as possible.
What are the next steps for your project?
Our capstone project is a practical example that we have explored only in theory and will not be pursuing further.
