ESG MOBILITY will be presenting the mobility platform metroSNAP at CES Las Vegas 2020 for the first time.
Since online commerce is booming and now also includes the fresh food sector, autonomous vehicles that swarm out and bring their products “just in time” to the customer without detours are the future. At metroSNAP, the chassis (skateboard) and superstructures (“PODs”) are interchangeable at any time. The chassis is planned as a fully autonomous service platform, while the bodies are adapted to the respective business case of any operator. The POD of the metroSNAP is placed in the city area, for example, to enable food purchases on the spot or to be able to deliver the mail to relevant traffic junctions. The metroSNAP is a logical first step in solving day-to-day tasks for people in a demand-oriented and decentralized manner.
With our connectivity solutions and an artificial intelligence suggestion system, we enable future operators to predict the highest revenue positions in the city.
Recommender system with artificial intelligence
The challenge of determining the right position in the city area, and constantly optimizing the position of the POD over the course of the day, can only be solved by machine learning. We have developed a recommender system for metroSNAP that predicts these optimal positions. The heart of the system is the so-called Deep Reinforcement Learning, which tracks the sales of PODs in the overall network as a target. Our approach does not require existing datasets to start, allowing new business models to be productive immediately.
Sophisticated end-to-end connectivity
Whether it’s the realization of customer features or complex analysis functions – a large amount of data from the metroSNAP vehicle fleet is needed. The future operators of a metroSNAP-based business model would like to be able to use comfortable features in an appealing fleet management tool. However, instead of collecting data centrally from the vehicle fleet for analysis or customer functions, very specific search jobs are sent to a vehicle and executed locally. Only the results find their way back to the backend. This approach is known as event-based data collection or campaign management. In the development phase, for example, we rely on our own connected data recorders to ensure data traceability and to recognize conspicuous behavior of the software in the vehicle.
New approach to the vehicle architecture
If, as with the metroSNAP, the most important vehicle components have to be exchangeable in live operation and completely new POD structures have to be adaptable to different operator models, this places unprecedented demands on the entire vehicle architecture. We develop logical function architectures as well as technical architectures in order to map the demanding customer requirements to the respective hardware. In order to be able to respond flexibly to changes and adjustments in the development process, we use state-of-the-art, agile methodologies.
Convince yourself of our expertise in machine learning, data driven business models and connectivity for mobility platforms at CES Las Vegas 2020: January 7th – 10th. Las Vegas Convention Center, North Hall, Stand 8516