Guide to Inertial Navigation System for Drones

An Inertial Navigation System, usually shortened to INS, is one of the most important technologies used in professional drones. It helps the aircraft understand how it is moving, how it is oriented and how its position is changing. This information is essential for stable flight, autonomous navigation, mapping, LiDAR surveying, BVLOS operations and flights in environments where GNSS signals may be unreliable or unavailable. At the centre of an INS is an Inertial Measurement Unit, or IMU. The IMU normally contains accelerometers and gyroscopes. Accelerometers measure linear movement, while gyroscopes measure rotational movement. By continuously processing this information, the drone can estimate its attitude, velocity and position. Modern drones rarely use inertial navigation on its own. Instead, INS is normally combined with GNSS, magnetometers, barometers, cameras, LiDAR, radar or other navigation sensors. Each system has different strengths and weaknesses, so combining them creates a more reliable navigation solution. ## **What Is an Inertial Navigation System?** An INS is a navigation system that estimates movement using inertial measurements. The IMU provides raw information about acceleration and rotation, while navigation software converts those measurements into a usable estimate of the aircraft’s position, speed and orientation. Unlike GNSS, inertial navigation does not need external satellite signals to measure short-term movement. This makes it especially valuable when GNSS is temporarily blocked, degraded or unavailable. However, inertial navigation has an important limitation. Small sensor errors accumulate over time. This means an INS can provide excellent short-term navigation, but its position estimate gradually becomes less accurate unless another sensor corrects it. ## **INS vs IMU** INS and IMU are often used interchangeably, but they are not the same thing. The IMU is the physical sensor package containing the accelerometers and gyroscopes. The INS is the complete navigation system that uses the IMU data to calculate position, velocity and attitude. A simple way to think about it is that the IMU measures motion, while the INS interprets that motion and turns it into navigation information. In many small drones, the IMU is built directly into the flight controller. Higher-end mapping and LiDAR systems may use a separate survey-grade INS because they need much greater positioning and orientation accuracy. ## **Accelerometers** Accelerometers measure linear acceleration along the drone’s three axes. These normally correspond to forward and backward movement, left and right movement, and vertical movement. The navigation system can integrate acceleration measurements over time to estimate changes in velocity. It can then integrate velocity to estimate changes in position. The problem is that even very small accelerometer errors can grow significantly after repeated integration. This is one of the main reasons inertial systems need regular correction from GNSS or another positioning source. ## **Gyroscopes** Gyroscopes measure how quickly the aircraft is rotating. They provide information about roll, pitch and yaw, which describe the orientation of the drone in three-dimensional space. Roll is movement around the aircraft’s longitudinal axis, pitch is movement around its lateral axis and yaw is rotation around its vertical axis. The autopilot constantly uses this information to understand how the aircraft is positioned and how it is changing orientation. Accurate gyro information is also extremely important for mapping. A small angular error can create a significant positioning error when a camera or LiDAR sensor is looking at the ground from altitude. ## **Attitude Estimation** Attitude describes the orientation of the drone in space. The INS continuously estimates roll, pitch and yaw using the IMU. This information is critical for stable flight because the flight controller needs to know whether the aircraft is level, banking or rotating. It is also important for geospatial applications because every camera image or LiDAR measurement needs to be associated with the correct sensor orientation. Professional systems may combine gyro information with magnetometers or dual-antenna GNSS to improve heading accuracy. ## **INS Drift** The biggest limitation of inertial navigation is drift. A tiny measurement error in a gyro or accelerometer can create a growing navigation error as time passes. For example, if an accelerometer reports a small amount of acceleration even when the aircraft is stationary, the navigation system will gradually calculate an incorrect velocity and position. The same effect occurs with small gyro biases. Higher-quality inertial sensors drift more slowly. This is one of the main differences between low-cost IMUs and higher-performance industrial or survey-grade INS systems. ## **GNSS and INS Integration** Most professional drones combine INS with GNSS. GNSS provides an absolute geographic position, while the INS provides very rapid short-term motion information. The two technologies complement each other extremely well. GNSS corrects long-term inertial drift, while the INS provides smooth updates between GNSS measurements. If GNSS is temporarily lost, the INS can continue estimating aircraft movement for a period of time. Once the satellite signal returns, the navigation system can correct the accumulated inertial error. ## **Sensor Fusion** Sensor fusion is the process of combining information from several navigation sensors to produce a stronger overall position estimate. A drone may use INS, GNSS, magnetometers, barometers, cameras and LiDAR simultaneously. Each sensor contributes different information. GNSS provides global position, barometers help with altitude, magnetometers provide heading information and cameras or LiDAR can provide local movement references. The navigation computer continuously compares these measurements and decides how much confidence to place in each one. This creates a more resilient navigation solution than relying on any single sensor. ## **Kalman Filtering** Many drone navigation systems use a Kalman filter or a related estimation algorithm. The filter predicts where the drone should be based on inertial information and then compares that prediction with measurements from GNSS and other sensors. When GNSS provides a new position, the filter can use it to correct the inertial estimate. If GNSS becomes unreliable, the filter can temporarily rely more heavily on other sensors. Modern drone autopilots frequently use Extended Kalman Filters because aircraft movement is highly dynamic and nonlinear. ## **GNSS-Denied Navigation** INS becomes especially important in environments where GNSS is unavailable. This can include warehouses, tunnels, mines, indoor industrial facilities, areas beneath bridges and locations surrounded by tall buildings. INS alone can maintain navigation only for a limited period because of drift. For this reason, GNSS-denied drones normally combine inertial navigation with cameras, LiDAR or radar. Visual-Inertial Odometry and LiDAR SLAM are examples of systems that use the IMU together with environmental sensing to maintain navigation without satellites. ## **Visual-Inertial Navigation** Visual-Inertial Navigation combines cameras with inertial data. The camera tracks visual features in the environment while the IMU measures rapid movement between frames. The two systems support each other. The IMU helps the computer predict how the camera is moving, while visual features provide corrections that reduce inertial drift. This technology is widely used in indoor drones, autonomous robotics and aircraft operating around structures where GNSS reception is poor. ## **LiDAR-Inertial Navigation** LiDAR can also be combined with INS. The LiDAR continuously scans the surrounding environment and creates a three-dimensional representation of nearby surfaces. The INS provides rapid