When Segway launched its AI-powered S90L e-scooter back in 2022, it mass produced 10,000 vehicles just one month after receiving factory orders.
Why the popularity?
A key factor was that the e-scooter came integrated with Segway Pilot – an ‘AI box’ offering lane and pedestrian detection for shared operators.
Since then, Segway has condensed all of these functions into its ‘Segway Pilot Edge’ add-on component which provides AI detection capabilities by attaching onto any of Segway’s shared e-scooters.
At Micromobility Europe, the vehicle solution provider launched its latest version: the Segway Pilot Lite. This offers all the same AI capabilities as the Segway Pilot and Segway Pilot Edge but at a more cost-effective price for operators.
For Segway, it’s not just about reducing the price but about reducing the barriers to AI technology and offering technological inclusivity for any-sized operator to enhance road and rider safety, city management and fleet operations.
The Segway Pilot Lite can attach to an even wider range of compatible models as either an integrated AI hardware module or an add-on component. With both its hardware and software entirely reconfigured, the Lite makes for Segway’s most cost-efficient AI hardware module, and it’s now undergoing mass production with client orders received.
“The Segway Pilot Lite answers the question that cities ask: why should we consider e-scooters to be safe?” Segway’s AI Product Manager Ric Ren tells Zag Daily.
“The answer is that improper riding can be regulated, which is essential to improving the overall safety of city transportation.
“As a provider of shared vehicle solutions, we have a responsibility to develop reliable, accessible, and intelligent solutions to protect road users. We’ve demonstrated how we can do this using AI.”
Cost-effective AI for all operators
How is it that the Segway Pilot Lite can boast all the same AI capabilities as the Pilot or Pilot Edge, yet be more cost-effective and lower the barriers to technology?
After years of AI model development and customer feedback, Segway found the answer by carefully reviewing its hardware to minimise redundancy, making it cheaper for operators to use.
The manufacturer observed how the Pilot was helping operators win tenders, with many clients requesting sample vehicles of the S90L to demonstrate to cities how safety is prioritised. By reconfiguring its hardware, the Lite lets smaller operators reap the same benefits at a more cost-effective price.
AI detections in brief
The Segway Pilot Lite features AI-powered lane and pedestrian detection.
For lane detection, the Segway Pilot Lite can identify changes in road surfaces to detect whether it’s a sidewalk or other lanes in 0.03 seconds. If it detects sidewalks where riding isn’t allowed, it automatically takes action based on preset programmes, which includes slowing down, giving prompts, and locally logging events.
To navigate busy streets, the AI box can automatically detect pedestrian dynamics and quantities. Again, due to its customised settings it can slow the vehicle down, give prompts, and locally log events when detected pedestrian count is higher than the operator-configured threshold.
“The Segway Pilot Lite analyses road elements and pedestrian dynamics to judge where the vehicle is driving and nearby people, and make sure it rides in line with the country’s legislation.”
By using computer vision – the Pilot Lite can determine the vehicle’s precise location and capture the real-world picture on the ground, which a GPS tracking system can’t do in an area with weak signal.
“The industry mainly relies on GPS tracking for location and positioning but its accuracy isn’t sufficient for road-level positioning,” Ric says. “We can use computer vision in the Segway Pilot Lite to empower shared micromobility and make scooters more road-user-friendly. It can locate and retrieve vehicles for operational efficiency and fleet management.”
The Lite’s dual-core central processing unit (CPU) is designed to ensure stable operations with fast and accurate detection thanks to its 1TOPS computing power tensor processing unit (TPU). It also features an anti-vibration design and has a full memory write cycle life of 3000 times. With a seven-year lifecycle with two hours of daily use, it aims to be tailored to the needs of various shared micromobility scenarios.
A wide-angle lens provides a 150॰ display field of view (DFOV) which can continuously monitor road conditions and pedestrian activity.
“Many cities hyperfixate on the possibility of illegal rider behaviour. We are using this technology to change their mind, and tell them that scooters can be safe and regulated.”
Lowering barriers to entry
The Segway Pilot Lite can either be integrated into the forehead of the vehicle or it can be added onto a compatible model as a versatile component.
Segway designed the Pilot Lite to be available in these two forms after its first AI box – the Segway Pilot – came integrated in the S90L, and the operator found existing clients with other e-scooter models also wanting these same AI performance technologies.
“Most of the time, when we think about AI for electric vehicles or for cars we think it’s very high-tech,” Ric says. “But for our market of shared micromobility, it’s a necessary safety requirement that will grow as the industry grows.
“We want AI to gradually standardise micromobility use and enhance its perception as a safe mode of transportation amongst the public. This will mean more cities adopt and endorse micromobility, and support its development even further.”
