Autonomous driving and ADAS

The car that sees the road, and brakes or steers on its own

Some new cars in India brake on their own when a crash is about to happen, or steer gently to keep the car in its lane. To do that, the car has to see the road. In a few American cities, taxis with no driver at all now carry paying passengers.

From assistance to self-driving. The industry sorts these systems by how much the car does:

Driver assistance

The driver drives, and the car helps: it brakes in an emergency, keeps the car in its lane, or holds a safe distance from the car in front. These features are called ADAS, advanced driver-assistance systems.

Partial automation

The car steers and brakes by itself in some conditions, such as on a highway, but the driver must watch and be ready to take over.

Full self-driving

The car drives with no one at the wheel, within a set area. Driverless taxis, called robotaxis, work this way.

The industry sorts these systems by how much the car does: driver assistance, where the car only helps; partial automation, where the car steers and brakes but the driver watches; and full self-driving, with no one at the wheel.

How a car sees. A car uses three kinds of sensor, often together:

Cameras

are cheap, and read colours, signs and lane markings. They struggle in darkness and heavy rain.

Radar

measures the distance and speed of objects, in any weather, but shows little detail.

Lidar

sends out laser pulses and builds a precise 3D picture of everything around the car. It costs more.

Software then does three jobs. It recognises what the sensors see, such as a car, a scooter or a person. It plans the car's path. And it controls the steering and brakes.

How a car sees and drives: cameras, radar and lidar feed software that recognises what they see, plans the car's path and controls the steering and brakes, with high-definition maps and simulation behind it.

Maps and simulation. A self-driving car also needs a high-definition map, precise enough to show each lane. (More on these maps in Market Segment 8.4, Mapping and location intelligence platforms, in Industry Vertical 8, Analytics and data providers.) And it must be tested far more than real roads allow. Rare and dangerous situations cannot be staged safely on real roads. So companies test their software in simulation first, over millions of virtual kilometres.

Where the work is. Carmakers buy driver assistance from suppliers. Robotaxi companies build their own systems. Most of the world's self-driving effort is in America and China, and the Indian work here is small. But the maps and simulation behind self-driving are built partly in India. India is also one of the world's largest markets for two-wheelers, so assistance systems for scooter and motorcycle riders are a natural place for Indian start-ups to begin.

Every company named here has posted software engineering jobs in India. Famous companies that don't actively hire software engineers in India are left out.

Keep reading, free

One more section is on this page: Self-driving, sensors, simulation and maps. Sign in to read it here, in full.

Continue with Google

Self-driving, sensors, simulation and maps

Self-driving, sensors, simulation and maps

Self-driving taxis

Waymo

owned by Alphabet, Google's parent company, has engineers in Bengaluru. It runs driverless taxis in several American cities. Waymo has the most job postings of any company covered mainly in this Market Segment.

Tools and sensors

Applied Intuition

with engineers in Bengaluru, sells simulation software, and the software that drives an autonomous vehicle. Carmakers and defence companies use it to build their own. (More on it in Market Segment 39.2, Drones and defence tech, in Industry Vertical 39, Aerospace, defence and space.)

Aeva

with engineers in Bengaluru, makes lidar that measures the speed of everything it sees, as well as its distance.

EdgeVerse India

makes perception software, which recognises what the sensors see, and rider-assistance systems for two-wheelers.

Virya Autonomous Technologies, in Bengaluru, is another company in this area.

Two companies covered mainly in other Market Segments also work here:

HERE Technologies

with engineers in Mumbai, makes the lane-level maps that assisted and self-driving cars depend on. HERE has more job postings than any company covered mainly in this Market Segment. (More on it in Market Segment 8.4, Mapping and location intelligence platforms, in Industry Vertical 8, Analytics and data providers.)

Rekise Marine

with engineers in Bengaluru, works at sea instead. It builds autonomous vessels that sail on the surface or under water. (More on it in Market Segment 39.2, Drones and defence tech, in Industry Vertical 39, Aerospace, defence and space.)

The AI model that drives.

Self-driving used to rely on rules written by hand. Now it relies on large AI models trained on millions of hours of driving, which learn to go from what the cameras see straight to steering and braking. This is the approach of Tesla and of Waymo's newer systems. Indian roads, with their mix of scooters, cattle and unmarked lanes, remain among the hardest in the world. Rider assistance for two-wheelers is where Indian start-ups are starting.

That is the car that sees the road: sensors that see, software that understands and drives, and the maps and simulation behind them. Most of the self-driving effort is abroad, but the maps and simulation are built partly in India.

Who they hire

Who these companies hire, and for what, is on What automotive and mobility companies hire for.

← Previous in Automotive and mobility

Automotive infotainment systems

Next in Automotive and mobility →

Connected car platforms