Luna Systems: Making ARAS Accessible

India's roads bear witness to a grim reality year after year. In 2023 alone, the nation recorded a staggering number of fatalities from road accidents, with the figure reaching nearly 1.73 lakh, according to data shared by states with the central government – a record high. This translates into an average of 474 lives lost every day, or one death occurring almost every three minutes. Alarmingly, the burden of this crisis falls disproportionately on riders of motorcycles and scooters.

These two-wheeler occupants accounted for approximately 44% of the fatalities in 2023, translating to over 76,000 lives lost. While advanced safety features, collectively known as Advanced Driver-Assistance Systems (ADAS), have become increasingly standard in four-wheeled vehicles globally, proving effective in reducing road deaths, their counterparts for twowheelers remain scarce.

ADAS technologies utilize sensors like cameras and radar to detect potential hazards or driver errors, alerting the driver or even intervening automatically to prevent collisions. Yet, motorcycles and scooters, particularly in high-volume, price-sensitive markets like India, have largely been left behind by this safety revolution.

Addressing this critical technology chasm is Luna Systems, a Dublin-based startup nearing its fifth year. Led by CEO Andrew Fleury, Luna is developing Advanced Rider-Assistance Systems (ARAS) specifically tailored for the unique dynamics and constraints of two-wheelers. With a strategic focus on India, the world's largest and arguably most dangerous market for motorcyclists, Luna aims to bring life-saving technology to millions of vulnerable road users.

India's Roads: A Deadly Landscape for Riders

The stark contrast between the advanced safety net available in modern cars and the exposure faced by motorcyclists is central to Luna Systems' mission. As Fleury noted, implementing technologies like computer vision and ADAS on two-wheelers is sometimes considered "a crazy idea," despite their proven life-saving potential in cars. The statistics, however, scream for a solution.

This need has given rise to Advanced Rider Assistance Systems (ARAS) – technologies specifically designed for the unique physics and operating environment of motorcycles and scooters . ARAS leverage sensors such as radar, cameras, and Inertial Measurement Units (IMUs), coupled with sophisticated algorithms, to offer safety functions like forward collision warnings, blind-spot detection, adaptive cruise control, and emergency braking support, all adapted for two-wheeled dynamics.

However, translating safety technology from cars to motorcycles is far from straightforward. Ujjwala Karle, Deputy Director at the Automotive Research Association of India (ARAI), emphasized this point, stating, "If you look into very conventional systems which will solve this, they will not work for two-wheelers".

Features common in cars, like lane-keeping assistance, are particularly challenging to implement effectively and affordably on motorcycles, especially for cost-sensitive markets like India. Engineering hurdles also exist; for instance, radarbased systems can struggle with the significant pitch, roll, and yaw movements inherent in motorcycle riding, potentially compromising detection accuracy, particularly during cornering or leaning. It was precisely this gap – the high vulnerability of riders and the lack of suitable technology – that spurred the founding of Luna Systems.

The company aims to specialize in bringing computer vision-based safety solutions to these road users who need them most. Luna enters this field as the global ARAS market begins to take shape, projected to grow from an estimated US$1.6 Billion in 2024 to US$2.6 Billion by 2030. While ADAS for cars is a mature segment, ARAS is still nascent, offering Luna the opportunity to establish itself early but also requiring it to navigate the uncertainties of an unproven market and evolving technology. 

Luna’s AI Leap

Luna's core technical strategy revolves around mastering computer vision and, crucially, low compute Artificial Intelligence (AI). The severe constraints of a typical motorcycle – limited space, power availability, and intense pressure on cost – necessitate AI models that are both powerful and exceptionally efficient. Fleury quantifies the challenge: Luna operates systems within the one-TOP (Trillion Operations Per Second) range of processing power, which is less compute capability than a standard smartphone.

This contrasts sharply with ADAS systems in cars, which often utilize hundreds of TOPS or more. “The challenge isn't just the cost—it's designing safety technology that works within the constraints of size, power, and processing limitations,” Fleury explained. Achieving sophisticated safety functions on such lean hardware is Luna's key technical differentiator. This low-compute capability enables a range of ARAS features Luna is developing. Some are adaptations from the four-wheeler world, such as headway monitoring (detecting if the rider is too close to the vehicle ahead) and forward collision warning (FCW), which alerts the rider if a potential impact is imminent.

Others are specifically localized for two-wheeler scenarios, like a warning for car doors opening into the rider's path and a reverse blind spot warning. Fleury highlights the reverse blind spot feature – alerting riders when they are entering another vehicle's blind spot – as potentially one of the most impactful, encouraging safer road positioning. Features like pedestrian detection and rear collision warnings (RCW), which monitor traffic approaching from behind, are also in development or demonstrable.

Perhaps Luna's most distinctive element is its focus beyond real-time alerts. Recognizing that ingrained habits contribute significantly to accidents, particularly in the context of widespread disregard for basic safety rules in markets like India, Luna emphasizes post-ride analytics aimed at behavior change. This "AI coach" concept leverages the camera data captured by the system to identify patterns of risky behavior, such as frequently tailgating or lingering in blind spots.

This feedback would be delivered to the rider via connected motorcycle applications, using Software Development Kits (SDKs) provided by Luna to the vehicle manufacturers (OEMs). OEMs could then integrate this coaching into their own apps and digital ecosystems, using their brand voice and potentially incentives or insurance links to encourage safer riding habits. Luna believes that directly addressing rider behavior in addition to providing alerts, could rapidly reduce two-wheeler fatalities.

This focus on camera-based AI and behavioral coaching represents a strategic choice. While radar systems are sometimes considered for ARAS, potentially offering initial cost advantages or better performance in adverse weather, they cannot provide the visual evidence necessary for Luna's AI coach.

By prioritizing cameras, Luna ties its technology directly to tackling the critical "human behavior" element identified by experts as a major cause of accidents in India. This positions their solution to potentially offer deeper, more sustained safety improvements, betting that this added value will outweigh the inherent limitations of cameras, such as reliance on clear visibility.

India First: The Proving Ground

Recognizing the unique confluence of market size, extreme need, and complex operating conditions, Luna has adopted an "India-first" development strategy. The company believes India, not Europe or North America, will lead the world in adopting ARAS technology. This conviction stems not only from the tragic fatality statistics but also from the sheer volume of two-wheelers used as primary daily transport and a perceived growing desire among both manufacturers and the public for enhanced safety.

The underlying logic is demanding: if Luna can make its ARAS reliable and affordable enough to work effectively amidst India's dense traffic, unpredictable road objects (including animals), and extreme price sensitivity, the technology should be robust enough for deployment anywhere else.

A cornerstone of this strategy is the localization of the AI models. Standard AI trained on Western datasets simply fails to comprehend the nuances of Indian roads. Girikumar Kumaresh, Principal Advisor for Road Safety at Bosch, highlighted this fundamental issue during an industry event , stating that even "Mature ADAS technology, when applied to Indian accident data, can only avoid 45-50% of accidents".

Examples abound: pedestrian detection systems trained to recognize Western gaits based on head and leg motion often fail when encountering traditional Indian attire that conceals these movements. Auto-rickshaws might be misclassified as cars, while certain agricultural vehicles like tractors, pulling mixers might go entirely undetected. In some cases, ADAS models trained according to UN regulations have even bizarrely classified buffaloes and cows as human beings.

ARAI's Ujjwala Karle reinforces this, noting the poor fit of conventional systems for Indian two-wheelers. To overcome this, Luna is investing significant effort in building an India-specific data foundation. The company is actively deploying data-capture motorcycles exclusively in India at present, gathering vast amounts of real-world road imagery. They have collaborated with the University of Pune to initiate this process, a relationship now evolving to include internships and advisory roles.

The objective is to train their AI algorithms on thousands, potentially millions, of images featuring uniquely Indian elements – from specific vehicle types like auto-rickshaws to prevalent animals – until the system can recognize and react to them reliably. This painstaking localization process, if successful, could provide Luna with a substantial advantage in the Indian market, creating a 'data moat' that would be difficult and time-consuming for competitors relying on generic datasets to replicate.

The Business Blueprint

Luna Systems is positioning itself strategically within the automotive supply chain as a software Tier 2 supplier. A critical hurdle for any technology targeting India's mass-market two-wheeler segment is cost. Luna is acutely aware of this, aiming for its ARAS system to add only around 5%, and "certainly less than 10%", to the overall price of the motorcycle or scooter. Achieving this assertive price point relies on leveraging the increasing trend of connectivity features being built into new two-wheelers.

As bikes incorporate more sophisticated compute chips for connectivity and digital displays, Luna plans to utilize these existing hardware platforms, requiring only a modest upgrade to enable their computer vision and ARAS functions. This "piggybacking" strategy is key to avoiding the prohibitive cost of adding entirely separate, highpower processing units. Funding this ambitious venture requires significant capital. Luna's journey began with seed funding from its founders. Since then, the company has secured support from Enterprise Ireland (a government agency that functions similarly to a venture capital firm with matching capabilities), angel investors, and EIT Urban Mobility.

To date, Luna has raised approximately €2 million. The company is now in the process of finalizing a new funding round, expecting to announce "a little bit more coming" shortly. Looking further ahead, Luna anticipates needing a larger funding round, potentially towards the end of 2025 or early 2026, to transition from the current innovation and development phase into full-scale commercialization and market deployment Luna's path to market follows a typical automotive supplier trajectory. The current focus is on delivering Proof of Concepts (POCs) to various OEMs, allowing them to evaluate the technology's capabilities and potential.

The next phase, expected to extend through 2025, involves collaborating closely with OEMs to integrate the ARAS into pre-production vehicle designs. Fleury acknowledges the process, stating they plan to start "humbly with individual design wins," but harbor ambitions for their technology to eventually become a standard feature across the market. Luna operates with a lean team, comprising around 12 engineers and 15 people in total.

Navigating Roadblocks

The regulatory landscape for ARAS on two-wheelers is still largely undefined. Luna's perspective is that, unlike car ADAS where regulations often mandate already established technologies, for nascent systems like ARAS, regulation will likely follow technological maturity and proven effectiveness. The onus, therefore, is on technology providers like Luna and their OEM partners to first demonstrate conclusively that these systems work reliably and deliver tangible safety benefits in real-world conditions.

To accelerate adoption and gather the necessary largescale data to prove effectiveness, one potential pathway suggested is government subsidies, similar to those used to boost the electric two-wheeler market in India. Fleury, added that it will enable a statistically relevant number of ARAS-equipped vehicles onto the roads faster, thereby generating the data needed to validate the technology's impact.

The success of Luna, and others pursuing similar goals, depends on a complex interplay of factors: continued technological advancement, astute business strategy, the willingness of OEMs and consumers to adopt new safety features, and potentially supportive government policies like subsidies or eventual regulations.