Tokyo’s Public Transportation Embraces AI for Improved Passenger Analytics
In collaboration with Idein, the Tokyo Metropolitan Government has initiated a pilot project aimed at harnessing the power of artificial intelligence to gauge passenger numbers on the Toei Subway, managed by the Tokyo Bureau of Transportation. The experiment was set at Jimbocho Station on the Toei Shinjuku Line, integrating Idein’s edge AI platform, named ‘Actcast,’ with cutting-edge AI cameras known as ‘ai cast.’
These AI cameras were employed to monitor and analyze both the number of passengers on board the trains and the passenger flow during boarding and alighting, by utilizing sophisticated AI applications. To enhance the accuracy of the AI, Tokyo transit staff also conducted manual counts for comparison and validation purposes.
Outcomes of the AI Passenger Counting Trial
The results confirmed exceptional precision in measuring the number of passengers in the areas near train doors. Yet, regions near train windows, which were fitted with UV-cut films, showed reduced accuracy in comparison. The team plans to hone the system further by repositioning the cameras and refining the AI models. Idein anticipates that, with improvements to the AI window area accuracy, simulations based on actual train structures could reach an operational precision exceeding 90%.
Moreover, the AI demonstrated a flawless track record in detecting people during boarding and disembarking. With sights set on future applications, additional testing during peak congestion times and specific vehicle types will continue.
Idein’s Actcast platform stands out as it enables the collection and application of real-world data through image and sound analysis using a compact and cost-efficient device as small as a credit card. Its proprietary acceleration technology ensures that even without AI model downsizing, leading-edge AI analysis can be performed, utilizing sensors equipped with cameras, microphones, and thermometers.
The images and footage captured by edge AI cameras are processed in real-time, with the assurance that no personal identifiers are transmitted to the cloud or retained on the camera itself, thereby mitigating data breach risks even in case of device theft. The processed data is stored exclusively as text, ensuring privacy and security.
Important Questions and Answers:
1. What was the purpose of the AI camera project on the Tokyo Subway?
The project aimed to use artificial intelligence to accurately count passengers on the Toei Subway lines, improving passenger flow analytics.
2. Where was the pilot project conducted?
The pilot project took place at Jimbocho Station on the Toei Shinjuku Line.
3. What is ‘Actcast’?
Actcast is Idein’s edge AI platform used in the AI cameras for the project.
4. How does the AI ensure privacy?
The AI cameras process images and footage in real time and store data as anonymized text, without transmitting or retaining any personal identifiers. This is crucial for protecting the privacy of the passengers while using public transportation.
Key Challenges or Controversies:
– Integration with Existing Infrastructure: Ensuring the AI system works seamlessly with the existing subway’s technology and infrastructure can be challenging.
– Accuracy in Different Conditions: The initial trials showed reduced accuracy in certain areas, possibly indicating a challenge in maintaining high performance in varying environmental conditions.
– Public Acceptance: The use of AI and cameras in public spaces can raise privacy concerns and require efforts to gain public trust.
– Data Security: While the project emphasizes data security, the use of real-time processing must continue to be fail-safe against potential cyber threats.
Advantages:
– Enhanced Precision: AI cameras can potentially offer more accurate passenger counts than manual counting, leading to better resource allocation and service planning.
– Real-time Data Analysis: Instant processing can help transit staff make quick decisions based on up-to-date information.
– Scalability: Once the system is perfected, it can be expanded to other stations and transportation modes easily.
– Privacy and Security: By only storing data as text and not transmitting personal identifiers, the system prioritizes passenger privacy.
Disadvantages:
– Initial Cost: The installation and integration of new AI cameras may involve significant investment.
– Potential Errors: The system may not be fully accurate in some situations, as indicated by the difficulties with UV-cut films.
– Maintenance and Upgrades: Continuous improvement and maintenance of the system will be necessary.
Related Links:
For more information on AI technology used in public transportation, here are some suggested links:
– Idein (Main domain for the company behind Actcast)
– Tokyo Bureau of Transportation (Main domain for the department managing Toei Subway)
It should be noted that the details about data security and privacy, as well as the Actcast platform specifics, add context to the potential for broader applications beyond just counting passengers, such as improving overall operational efficiency and passenger experience on mass transit systems. The success of this pilot could lead to broader adaptability into other areas of public transportation or even different sectors where crowd monitoring and management is necessary.