More sensors or AI?

With continually improving artificial intelligence (AI) technology, system designers will face the design tradeoff of adding more sensors to achieve specific design objectives or using AI instead. A good example is a smart elevator. Today’s elevators have several sensors in their design.

One of the key features of smart elevators starts with traffic analysis and optimization. This can be accomplished with the addition of vision sensing to a new elevator design.

In a vision-based intelligent multi-elevator system, cameras are installed in the elevators and on each floor, with each camera connected to the elevator control system. Using the installed sensors, the elevator control system’s elevator people statistics module counts the number of people waiting for the elevator on each floor and the number of people in the elevator. Knowing the number of people waiting and the number of people already in the elevator, the system can control the stops and availability of the elevator.

However, adding these vision sensors can be costly. Through AI, smart elevators can analyze traffic patterns to optimize elevator car movement to reduce wait times and congestion. This is especially important in an infrequent situation, such as a cruise ship. Passengers need to easily learn how to use a new elevator system and adjust to its nuances quickly.

One example of implementing artificial intelligence to design a smart elevator system for marine applications is provided by Kone Corporation. Using its KONE TrafCal elevator traffic analysis, the multi-elevator system receives an assessment of the passenger flow together with an analysis of different situations, such as embarkation, disembarkation, and normal everyday operation at sea. This analysis provides a solution for the best possible flow of people onboard.

In addition, the company’s hybrid destination control system (DCS) significantly increases the handling capacity of elevators on board passenger vessels, particularly during peak periods. This is accomplished by grouping people traveling to the same destination into the same elevator. Instead of calling an elevator and selecting the desired floor after entering the elevator, this system points passengers to a specific elevator that has been selected because of its availability.

Figure 1. After selecting their destination, passengers are directed by their initial elevator’s display to a specific elevator to reduce wait time and avoid an already packed elevator arriving for them.

The system already detects passenger load, so it uses this information to send an elevator with available space to the floor where riders have requested an elevator to a specific floor. This avoids the frustrating wait for an elevator during rush times, only to have it arrive packed with no room for more riders. The KONE hybrid DCS combines all the performance advantages of a traditional DCS with the ease of use of a standard collective control system.

Figure 2. In the smart elevator, passengers are notified of their destination relative to the current floor (deck).

As noted in a previous blog, recognizing and meeting the requirements of industry standards and government regulations is essential for many different applications. For marine applications, Kone complies with classification societies registered under the International Association of Classification Societies (IACS), for example, Achilles JQS (Joint Qualification System), DNV, Lloyds, and RINA.

The tradeoff between adding more sensors and greater use of AI is a situation both sensor and system designers will need to address in many applications, especially those applications with several sensors already designed into existing systems.