KartaCam 2

LOW KO WEE (KW): We initially started out using third-party map services and bought some off-the-shelf cameras for street-level imagery. We had spreadsheets to manage the allocation of drivers to particular areas and streets for the collection of imagery. It was quite labour intensive. The street images were then used to create maps in order to improve the service quality of both the Grab superapp and our driver app.

The collection of street images was previously done with helmet-mounted cameras. Some of the roads in Southeast Asia don’t have smooth paving, and bumps resulted in blurry images. That means you’d have to go and record it again. Also, the drivers would have to manually remove the SD (secure digital) cards from the cameras and bring them to the operations team to transfer the files. The files would then need to be uploaded to the cloud for processing.

ELLEN CHEN (EC): We realised that using better and purpose-built hardware and software was crucial. KartaCam 2 has made significant improvements to the workflow and the efficiency of map creation. We proactively developed our own hardware and software and used crowdsourcing for map creation. We could do this because we have our own network of two-wheeled vehicle drivers who can access the small alleys and blocked roads that you often find in Southeast Asia.

KW: We innovated in multiple areas and the true distinction of KartaCam 2 lies in its integrated map-making solution. There are three critical elements. Firstly, there’s the hardware for high-resolution images and geospatial data collection. Secondly, there are the intuitive software interfaces designed for operational efficiency. Thirdly, there’s the seamless system integration, which allows an uninterrupted data flow throughout the mapping workflow.

EC: Grab’s proprietary navigation software is a crucial part of the solution, because it handles route guidance for drivers. It removes the psychological burden of figuring out the shortest route when collecting street imagery. This means they can concentrate fully on driving without the concern of unfamiliar roads.

KW: Beyond technical capabilities, innovation really comes down to the people on the team as well. Our IoT (Internet of Things) team has strong capabilities in the fields of mechanical engineering, electronics, and software development. All of this expertise came together, enabling us to optimise the product size and affordability, while keeping it super powerful.

KW: There were many! From a design perspective, the goal was really to simplify the image-collection workflow and free up the driver’s mental bandwidth while collecting street imagery. The considerations went far beyond the user interface. The team needed to understand the user-journey pain points during mapping.

For instance, the camera design needed to be flexible enough to fit most motorbikes and robust enough when mounted to endure potholes without being damaged. It also needed to be specifically engineered to operate reliably in hot conditions for extended periods without failure. These were all challenges that had to be overcome.

KW: From the start, the goal was always to solve user problems and deliver value. In terms of design, we have a core belief in turning complexity into simpler and better experiences. This doesn’t mean reduction. It means starting with a deep understanding of systems, user needs, and technical constraints. After that, we carefully remove anything that doesn’t serve the user.

EC: We always emphasise the seamless integration of hardware and software because this is how we achieve an intuitive and frictionless user experience across our devices and mobile application. So, across KartaCam 2 and the mobile application, there’s consistency in the user-interface behaviour and visual elements like icons.

KW: The magic often happens when designers are not working in silos, but in discussion. We did a lot of problem solving with the designers and engineers in the early phase of the product development. There can be some tension, because everyone brings expertise in different areas to the table, but we were able to find alignment on the best possible outcome.

EC: During the product planning phase, one of our key goals was to gather user feedback from testing on the road as early as possible. We created working prototypes for drivers to interact with and give feedback on. This iterative approach allowed us to refine the design based on real user input from around Southeast Asia before we finalised the design.

KW: Working cross-functionally wasn’t new to Grab, but there was a higher level of complexity in the testing and iteration of KartaCam 2 because of the combination of hardware and software. We conducted regular design jams and reviews, multiple rounds of usability testing, as well as the on-the-ground immersion with the drivers.

From the start, we set a high bar for the team: create the smallest product in its category and go beyond merely solving problems to deliver something that would be delightful to use. The process reinforced why our designers should view hardware and software as a unified product. If we lack a holistic perspective, blind spots can emerge late in development, limiting our ability to make meaningful hardware changes after we receive user feedback.

KW: Regarding the hardware, we made some important discoveries while testing an earlier version of the mounting system. A lot of our tanks under the seat rather than at the front of the bike. That meant our drivers spent a considerable amount of time removing the camera mount when refuelling only to set it up again.

We also discovered that the weight of the old mount made it difficult to manoeuvre through narrow alleys. Side mirrors were getting damaged. People also shared how the heat  the camera to shut down, leading to driver downtime. We thus needed to integrate a means of cooling the camera.

EC: Our female drivers have an average height of 150 cm, and we learned that they had difficulties viewing the camera and adjusting the shooting angle after mounting it on their bikes. We also learned that drivers preferred to dismount the camera when they entered indoor locations because of the risk of theft. These discoveries all resulted in refinements to the design.

KW: Grab drivers do not have to buy the cameras from us if they sign up as mappers. Grab provides all the necessary tools and training to help them map Southeast Asia. Some drivers prefer mapping over deliveries because it allows them to fully manage their own schedule while still earning an expected income.

KW: The new Tengah town is huge, with 42,000 homes. The residents were making reports about transport and delivery issues. For example, when they ordered a Grab car, the driver would park somewhere else and they’d have to walk to a pick-up point. The issue was that the roads were unmapped and this was impacting people’s daily lives. The GrabMaps team assisted by mapping the new estate to help address these challenges.

EC: Roads in Southeast Asia can change quite dramatically. KartaCam 2 is not only used for the initial creation of the map, but also its long-term maintenance. The street imagery we collect can be used for other purposes, such as enabling government bodies to locate potholes through an online portal and enabling workers to be dispatched to the correct location to repair the road.

KartaCam 2 offers a lot of versatility in terms of its use. The technology within KartaCam 2 could be manifested in other products such as as KartaDashCam, focusing on safety. KartaCam 2 itself can be upgraded with LiDAR capabilities for high-definition mapping.

KW: With better maps, we’ve seen significant improvements in the earnings of transport and delivery drivers because the platform is much more efficient. We’ve achieved better routing for drivers when picking up passengers or delivering food, which helps reduce road congestion and results in fuel savings. Also, Grab’s consumer app can now offer drivers a diverse selection of points of interest as destinations for bookings.

KW: Similar to how Grab used to rely on third-party mapping, our clients use it to build maps for their own platforms. It allows them to have better map accuracy, which can lead to faster and more cost-effective real-time fulfilment.