Advanced Rider Assistance Systems (ARAS) & Micromobility

The Evolution of Scooter Safety Technology

The ambition to address a seemingly straightforward issue – preventing shared electric scooters from being ridden on sidewalks – has spurred significant technological advancements within the micromobility sector.

Initially reliant on geofencing, scooter companies are now actively deploying technologies akin to those found in advanced driver assistance systems (ADAS), commonly utilized in automobiles, in a bid to gain a competitive advantage.

Implementation by Leading Operators

Companies such as Spin, Voi, Zipp, Bird, and Superpedestrian are making substantial investments in both camera-based and location-based technologies.

These systems are designed to identify and even rectify unsafe rider actions, with some capable of automatically reducing scooter speed or bringing the vehicle to a complete stop when sidewalk riding is detected.

Addressing Urban Concerns

Sidewalk scooter usage and improper parking represent a major concern for municipalities.

This issue frequently fuels complaints from residents resistant to change, particularly when it introduces potential safety hazards like tripping risks.

Companies are proactively attempting to resolve this problem through technology that places the responsibility for rider conduct on the operators themselves.

This may ultimately lead to cities mandating the implementation of ADAS-like technology for all scooter operators.

Cost-Effectiveness and Practicality

Scooter ADAS is emerging as a particularly feasible and economical solution for cities aiming to curb undesirable rider behavior.

It presents a more affordable alternative to direct policing of riders or the costly development of dedicated cycling infrastructure.

The Underlying Need for Infrastructure

“This technology stems from a necessity for protected bike lanes,” explains Dmitry Shevelenko, co-founder and president of Tortoise, a micromobility positioning service.

“It arises in situations where riders are compelled to engage in behaviors that may inconvenience others, simply due to a lack of alternative routes.”

Long-Term Solutions and Operator Responsibility

While the long-term solution involves constructing bike lanes and designated scooter parking areas, operators must currently demonstrate to local governments that micromobility is both safe and beneficial.

“Until cities establish dedicated infrastructure for emerging transportation methods, it’s crucial to leverage technology to prevent conflicts,” states Alex Nesic, co-founder and chief business officer of Drover AI, a computer vision startup specializing in scooter ADAS.

“Our goal is to facilitate the industry’s continued growth and maturity.”

Street Views and Satellite Views in Micromobility

Drover AI collaborates with Spin, whereas Luna, a fellow computer vision firm, partners with Voi and Zipp to integrate cameras, sensors, and a microprocessor onto scooters. This integration enables the detection of lanes, sidewalks, pedestrians, and surrounding environmental features.

According to Nesic, this functions as an additional IoT module mounted on the scooter’s stem, performing computations through sensor fusion directly on the device. “The Drover system distinguishes between sidewalks, streets, and bike lanes as distinct categories, continuously monitoring location and transitions between them.” This data is relayed to the scooter’s onboard systems in real-time, allowing for latency of under one second for implementing onboard alerts or speed adjustments.

Spin is currently trialing this technology with numerous devices across Milwaukee, Seattle, Miami, and Santa Monica. Testing includes the capability to automatically decelerate scooters. In other locations, Spin scooters emit an audible signal to alert riders to reduce speed or dismount from sidewalks.

Furthermore, the scooters transmit data to Spin via the cloud, triggering a notification on the rider’s application shortly after, explaining the reason for the audible alert or deceleration. Luna’s technology also possesses the ability to halt the scooter, a feature Voi is evaluating for its Northampton pilot program, alongside the audible warning.

Superpedestrian is adopting a distinct strategy for sensor fusion to monitor rider conduct. The company recently acquired Navmatic, a firm specializing in real-time vehicle localization and movement correction for micromobility operators, and is deploying a safety system known as Pedestrian Defense.

Rather than adding external hardware to scooters, Navmatic utilizes software operating on the scooter’s existing operating system, which also contains city maps. During operation, data from sensors—including vehicle direction, wheel speed, and motion—is combined with GPS data. This precise calculation of vehicle location and movement allows Superpedestrian to identify sidewalk riding, improper parking, and aggressive driving behaviors.

Bird is pursuing a comparable path to Superpedestrian, integrating a sensor fusion chip into its newer scooter models. The company, which initiated pilot programs in Milwaukee and is preparing to test in San Diego, asserts that this solution will be less susceptible to errors and more dependable across diverse terrains.

Both Bird and Superpedestrian have prioritized location-based ADAS over camera-based systems due to cost-effectiveness and scalability. They believe this approach leverages currently available and viable technology, while computer vision remains too costly and underdeveloped.

“It’s encouraging to witness a collective push for enhanced pedestrian safety within the industry,” stated Paul White, senior director of public affairs at Superpedestrian, to TechCrunch. “However, safety technology is only effective when deployed across all scooters. The most practical and enduring solutions are often the most economical. We’ve developed a method to improve rider safety without incorporating fragile, expensive, or difficult-to-integrate components. We can implement our Pedestrian Defense system fleet-wide, without increasing prices for customers or compromising labor, maintenance, or service expansion in underserved communities.”

Nesic affirms that Drover AI’s systems achieve 95% accuracy and do not rely on GPS, which can be inconsistent. “Our edge-based system maintains functionality even in areas with limited connectivity,” Nesic explained. “GPS-dependent systems require precise infrastructure knowledge, including accurate or high-definition mapping. Cities evolve rapidly, meaning a sidewalk today may not exist tomorrow. Our system doesn’t necessitate this level of ground truth, making it more adaptable to changing environments.”

Superpedestrian utilizes GPS, but Navmatic’s technology can transition to alternative sensors to maintain accurate positioning when GPS signals are unreliable, according to a company representative. Industry analysts suggest the Navmatic acquisition was primarily for its personnel, as Superpedestrian is known for recruiting skilled professionals, rather than the IP itself.

The spokesperson also indicated that Navmatic’s sensor fusion technology would provide a foundation for potential future computer vision integration, should the company choose to pursue that direction.

Investment Trends in Scooter ADAS Technology

While substantial capital hasn't yet flowed into the Advanced Driver-Assistance Systems (ADAS) sector for scooters, it’s a relatively nascent field compared to the broader shared scooter market. Therefore, the long-term investment landscape will become clearer as practical applications are further developed and assessed.

Luna secured €400,000 ($470,000) in seed funding in March, though it is highly probable that Voi and Zipp are currently compensating the company for the services it provides.

Voi recently completed a $45 million funding round in August. These funds are earmarked for continued research and development, specifically focusing on technologies designed to deter sidewalk riding and improper parking, potentially through internal development efforts.

Drover’s PathPilot Development and Funding

According to Nesic, the development of Drover’s PathPilot technology spanned approximately two years prior to the company's establishment. To date, the startup has secured around $900,000 in funding, valuing the company at $5 million pre-money.

“Our initial growth was fueled by a significant prepayment received from Spin upon signing our first contract,” Nesic explained. “This allowed us to expand our customer base internationally.”

Drover has now secured orders that will lead to positive Earnings Before Interest, Taxes, Depreciation, and Amortization (EBITDA) upon fulfillment. The company is currently seeking additional capital to address increasing demand from its clientele, though specific details remain undisclosed.

Superpedestrian’s Acquisition of Navmatic

The financial specifics of Superpedestrian’s acquisition were not publicly released. However, sources close to the deal indicate that Navmatic was purchased for approximately $4 million.

The Financial Implications for Scooter Operators

The ability of scooter companies to operate is heavily dependent on securing permits from city authorities, and this pressure is likely to intensify as competition increases and cities seek favorable solutions. Currently, it remains uncertain whether integrating advanced technologies will translate into increased profitability through permit acquisition, or if the associated costs will diminish potential earnings.

Significant operational expenses for shared micromobility services include overhead – encompassing insurance, fees, and rebalancing – alongside vehicle recharging and maintenance. Integrating computer vision modules into each scooter is estimated to add $100 to $200 in hardware costs per unit, according to industry sources. Furthermore, each additional component introduces a potential point of failure, potentially increasing maintenance demands.

Despite these initial investments, Nesic anticipates that the technology will ultimately generate cost savings for clients. For instance, scooters parked within garages create GPS blind spots, increasing the risk of theft or requiring costly recovery efforts. A camera-based system offers the capability to pinpoint a scooter’s location using visual data.

“Addressing parking violations is another key benefit,” Nesic explained. “Onboard cameras provide verifiable photographic evidence – regardless of whether the parking is compliant – at the conclusion of every ride. This empowers cities to manage scooter deployments more effectively, allows operators to proactively avoid penalties, and encourages users to park responsibly.”

The financial impact of parking fines can be substantial. San Francisco’s Municipal Transportation Agency (MTA) has documented numerous instances of improperly parked scooters, each incurring a $100 fine. Since November 2020, Lime has faced over 3,800 such violations, Scoot approximately 3,000, and Spin around 1,700.

Nesic suggests that incorporating Advanced Driver-Assistance Systems (ADAS) into scooters could potentially reduce insurance premiums for operators. However, insurance experts believe the situation is more complex.

“From an insurance standpoint, any factor that lowers both the likelihood and severity of a claim directly impacts the risk associated with insuring an event,” stated Mike Shim, co-founder and CEO of DigiSure, a digital insurance provider. “Therefore, if the insured event – property damage or liability resulting from an accident involving the scooter renter – can be mitigated through onboard technology, it theoretically should lead to a more favorable insurance rate.”

However, insurance rate adjustments are not immediate; they require substantial claims data analysis over extended periods.

“The challenge with this actuarial approach is that these models evolve slowly, typically over decades,” Shim continued. “By that point, we may well have fully autonomous aerial vehicles! This offers limited immediate benefit to current scooter companies.”

Future Investments in Scooter ADAS

A prevailing viewpoint suggests that ADAS (Advanced Driver-Assistance Systems) for scooters represents a safety feature with limited operational impact. It’s often seen as an additional hurdle for new companies seeking entry into the market, increasing expansion costs. However, considering the urgency of the climate crisis, restricting zero-emission transportation options could prove to be a detrimental misallocation of resources, yet this trend may persist.

Shevelenko noted that regulatory patterns often establish a momentum of expectations within a sector, making reversal difficult. Cities frequently adopt regulations from one another, potentially solidifying these requirements over time. “After two decades, these stipulations may become ingrained, with the original rationale or origins largely forgotten.”

The automotive industry offers a potential parallel for the trajectory of scooter ADAS. In 2019, the European Commission proposed mandating speed-limiting devices on all new vehicles in Europe and the U.K. by 2022. While the necessary technology was available, political factors previously hindered its implementation.

Although the U.S. typically lags behind Europe in adopting such measures, the possibility remains. A current bipartisan $1 trillion infrastructure bill under Senate consideration includes a provision for driver monitoring systems designed to detect and prevent drunk driving, potentially even through intervention.

A recent incident involving an intoxicated individual operating a personal e-scooter on a Birmingham motorway underscores the distinction between geofenced, shared scooters and privately owned models with unrestricted access. Should ADAS become prevalent in the shared scooter sector, it’s logical to anticipate its adoption in retail e-scooters and e-bikes, mirroring how consumer drones utilize geofencing to prevent operation in prohibited areas.

Conversely, the technology could face resistance from riders who perceive it as overly restrictive. Should this occur, companies such as Drover AI and Luna could consolidate their data and expertise to enhance the technology used in autonomous delivery robots. The future remains uncertain.