High-performance Inertial Navigation Systems for self-driving cars

Self-driving cars are vehicles equipped with sophisticated systems that enable them to navigate and control themselves without human intervention. These vehicles use a combination of autonomous driving sensors and algorithms to perceive their environment, make decisions, and perform self driving tasks. The goal is to achieve full autonomy, where the vehicle can handle all aspects of driving safely and efficiently.

Self-driving cars rely on an array of key technologies to perceive their surroundings. These include:

• GNSS (Global Navigation Satellite System): to get the real-time updates on the self-driving car position, speed, and direction.
• INS (Inertial Navigation Systems): to maintain accuracy in case of GNSS signal outages. It provides real-time updates on the self driving car position, speed, and direction.
• LiDAR (Light Detection and Ranging): use of laser beams to create a detailed 3D map of the vehicle’s environment. This technology helps the car detect and measure objects around it, including other vehicles, pedestrians, and road signs.
• Radar (Radio Detection and Ranging): use of radio waves to detect the speed, distance, and direction of objects. Radar is particularly useful in adverse weather conditions and for detecting objects at longer ranges.
• Cameras: to capture visual information about the vehicle’s environment, including lane markings, traffic signals, and road signs. They are essential for interpreting complex visual cues and making decisions based on visual data.

ADAS (Advanced Driver Assistance Systems) enhances driving safety by providing features like lane-keeping, adaptive cruise control, and automatic braking, but requires active driver supervision. In contrast, self-driving cars, equipped with autonomous driving systems, aim to fully automate vehicle operation without human intervention.

While ADAS supports drivers by assisting with tasks and improving safety, self-driving cars are designed to handle all aspects of autonomous driving, from navigation to decision-making, offering a higher level of automation (SAE levels) and convenience. ADAS characteristics or features are attributed to SAE levels below 3 and self-driving cars as such correspond to minimum level 4.

GPS (Global Positioning System) works by using a constellation of satellites, precise timing, and trilateration to determine your position anywhere on Earth.

Here’s the simplest clear explanation:

1 – Satellites broadcast signals

About 30 GPS satellites orbit Earth, each continuously transmitting:
– Its exact position in space
– The exact time the signal was sent (using atomic clocks)

These signals travel at the speed of light.

2 – Your receiver measures travel time

A GPS receiver (in your phone, drone, INS, etc.) picks up signals from multiple satellites.

By measuring how long each signal took to arrive, it computes the distance:

               distance = speed of light × travel time

3 – Trilateration computes your location

To find your position, the receiver uses trilateration (not triangulation):

• With 1 satellite → you could be anywhere on a sphere
• With 2 satellites → circles intersect
• With 3 satellites → two possible points
• With 4 satellites → your exact 3D position + clock correction

Your receiver doesn’t have an atomic clock, so the 4th satellite is needed to solve timing errors.

4 – Corrections improve accuracy

Raw GPS has errors from:

• Atmosphere (ionosphere, troposphere)
• Satellite clock drift
• Orbit prediction errors
• Multipath reflections (signals bouncing off buildings)

To improve accuracy:

• SBAS (e.g., WAAS, EGNOS) provides real-time corrections
• RTK and PPP techniques correct errors down to centimeter-level
• INS coupling (IMU + GPS) smooths and bridges gaps during signal loss

6 – Final output

The receiver combines all data to estimate:

• Latitude
• Longitude
• Altitude
• Velocity
• Precise time

Modern GPS receivers do this dozens or hundreds of times per second.

An INS (Inertial Navigation System) is a self-contained navigation solution that determines a platform’s position, orientation, and velocity using only inertial sensors—typically:

• Accelerometers (measure linear acceleration)
• Gyroscopes (measure angular rotation)

How it Works ?

Gyroscopes track how the platform rotates (roll, pitch, yaw). Accelerometers measure movement along three axes. A navigation filter (usually a Kalman filter) integrates these measurements over time to compute:

• Position (x, y, z)
• Velocity
• Attitude (orientation)

Key Characteristics

1. Fully autonomous: No external signals needed to operate
2. High update rate: Often hundreds or thousands of measurements per second
3. Operates in any environment: Works underground, underwater, indoors, and in GPS-denied environments
4. Precision depends on sensor grade: Ranges from consumer-grade IMUs to tactical and navigation-grade INS

Common Applications

• Aerospace & defense: missiles, UAVs, loitering munitions, armored vehicles
• Marine: AUVs, USVs, ships, hydrography systems
• Land robotics: autonomous vehicles, SLAM, AGVs
• Surveying & mapping: mobile mapping systems, LiDAR
• Industrial: stabilization, motion tracking