Building semi-autonomous e-scooters for the future

Bengaluru-based startup Flo Mobility is building semi-autonomous e-scooters by leveraging computer vision, AI for obstacle identification, robotics and the Internet of Things (IoT) sub-systems. Manesh Jain, Founder & CEO, Flo Mobility shares that they are targeting five European cities (Berlin, Stockholm, Paris, Amsterdam and London), three in the Middle East for pilot projects, and have have identified 200+ large campuses in India where the solution will be useful for commuting within the premises

The story of your beginning, how did you come up with the concept of semi-autonomous e-scooters ?

Manesh Jain, Founder & CEO, Flo Mobility

We launched our operations in July 2020. Given co-founder Angad’s extensive experience in the electrical vehicle domain and other co-founder Manesh’s keen interest in urban mobility, we got really excited about the evolution of micro-mobility and role electric vehicles can play in offering an eco-friendly, convenient and easy mode of commute for the last mile.

So we started our research and realised that urban traffic congestion and the last-mile commute is a big concern for authorities all over the world. Several companies are trying to solve this problem by offering electric scooters for ride-share. Not only are these scooters environment friendly, but they also help riders save commute time. But these scooters clutter the pathways and obstruct pedestrians when not parked properly. Also, they need to be charged timely to ensure that they are available for booking. Moreover, the rider needs to locate and walk to the vehicle to start the journey and leave at a proper parking station to end the journey which is a hassle.

So we decided to look for a solution that can address these challenges and eventually came up with the idea of building autonomous 2 wheelers.

Which cutting edge technologies are you using to develop these e-scooters ?

We have developed a remote repositioning solution for electric 2 wheelers using a IOT and hardware stack consisting of cameras, sensors, sonar, edge processors and controllers.
Core technology: Teleoperations, Computer vision, Embedded systems
Tech Framework: Robot operating system (ROS), webRTC tunnelling, OpenCV based Object Detection, Edge video streaming and action relay,
IP: Eventually we will have 3D maps of the roads of the localities we are working in.

Do you believe in the future remote teleported vehicle movements will become common in the country ? In which cities / towns do you foresee this trend ?

Teleoperation is the first step towards autonomous driving. We envision a world soon where autonomous vehicles will take precedence over manual driving. Tesla, Waymo, Starship and multiple other companies have shown it that autonomous driving is safer and cost effective than manual driving. Germany has already stated that it will allow on road testing of autonomous vehicles by 2023. San Francisco and Los Angeles have been experimenting with autonomous delivery robots for quite some time now. Starship completed a 1 millionth trip in january 2021 via autonomous delivery robots and Waymo has driven over 20 million miles autonomously on public roads across 25 U.S. cities.

Autonomous and teleoperations will be first seen in the US and Europe where road rules and road signs are pretty clear. But soon with enough data, we will see autonomous driving vehicles everywhere.

Which technologies have you incorporated in your FLO-TOK kit that can be installed in existing scooters ? Please explain in detail about your teleportations software.

There are two major components to our technology; hardware and software. Hardware involve the sensors and mechanical interventions on the scooter to drive the vehicle around. Our hardware kit includes 3 cameras, 3 ultrasonic sensors, a couple of motors and supplementary electronics.
On the software side, we are using advanced matching algorithms to match the teleoperators and teletrips based on various factors.

Our path planning and object detection algorithm triangulates the data provided by all the sensors and generates intelligence to navigate around the object and keep moving towards the destination. For additional safety and redundancy, fall back algorithms are created that act as emergency triggers for actuators in case primary decision mapping is delayed or inaccurate.

While the on-board decision making is orchestrated using edge computing, the tele-operator at backend administers the entire process and guides the vehicle providing an auxiliary decision making mechanism that works in tandem with the on-board decision making. Operators can supersede the on-board decisions to handle corner cases and extraordinary situations.

Your focus on innovation.
Iterative innovation is the main backbone of the company. We believe in making all the components as efficient as possible and therefore need continuous experimentation and innovation. Current focus areas for innovation being mechanical attachments to the scooter, a smart processor to compute data on the edge, algorithms to detect potholes, speed bumps, entry/exit to footpath and bicycle lanes, etc.

How do you plan to popularise your simulation training systems ?

Our focus is not to popularise the simulation training systems but rather the autonomous tech itself. We plan to approach urban planning authorities and policy think tanks and present the benefits of electric 2 wheelers and how autonomous tech can help improve the adoption of electric vehicles. Our immediate implementation area would be last mile commute, where we will offer our tech to micromobility companies who deploy fleets for riders to rent vehicles on demand. We are also planning to conduct pilots in large business parks and residential complexes to showcase the capability, comfort and convenience that our tech brings to urban commutes.

Your plans for the future, including timelines for the launch, etc.

We intend to scale our business to 1000 vehicles by the end of 2021. This would include both its own fleet as well as 3rd party fleet on which we would retrofit our hardware. The focus areas for deployment over next year are university campuses, business parks, residential complexes, resorts, tourists spots and other similar closed premises. We are targeting five European cities Berlin, Stockholm, Paris, Amsterdam and London as well as three cities in middle east Dubai, Abu Dhabi and Sharjah for pilot projects. In India, we have identified 200+ large campuses where our solution will be useful for commuting within the premises.

Any other significant factor ?

There are more than 250k electric scooters deployed in several large cities across the globe. Though they provide a very convenient and fast mode of commute, several city authorities have banned them or have put hefty penalties due to non-compliance of parking rules. Riders often leave the scooter on the sideways, obstructing pedestrian paths. So on one side city authorities want to promote scooter but at the same time, the menace they create when not managed properly is a big deterrent in doing so.

Artificial IntelligenceComputer VisionFlo MobilityIOT
Comments (0)
Add Comment