3.3. Sensor Calibration and Noise Modeling

🎯 Why Calibration is Necessary Sensor readings can be inaccurate due to systematic errors or misalignment. Calibration ensures that the data accurately reflects real-world measurements.

Calibration is crucial in robotics because sensors often produce readings that are not accurate. This inaccuracy can arise from various issues like systematic errors or the physical misalignment of sensors. Calibration aims to adjust these readings to ensure that they represent the real-world conditions as closely as possible.

Imagine using a scale to weigh yourself, but the scale shows you weigh 5 kg less than your actual weight. Before you can trust the weights, you must adjust the scale to reflect your true weight. Similarly, calibrating sensors ensures they provide accurate readings.

Types of Calibration: ● Intrinsic Calibration: Corrects internal distortions (e.g., lens distortion in cameras). ● Extrinsic Calibration: Aligns one sensor’s frame with another (e.g., camera-to-LiDAR alignment). ● Temporal Calibration: Synchronizes sensors running at different frequencies. 🧪 Example: Aligning the depth image of a LiDAR with RGB images for accurate fusion.

There are three main types of calibration:

1. Intrinsic Calibration focuses on correcting errors within a single sensor, like correcting lens distortion in a camera image to ensure it captures true shapes and sizes.
2. Extrinsic Calibration is essential when multiple sensors are used together, making sure they are aligned correctly, as in the case of aligning images from a camera with data from a LiDAR sensor.
3. Temporal Calibration involves synchronizing data from sensors that operate at different rates to ensure that all readings refer to the same time frame. This is important, for example, when comparing data from a fast camera to a slower LiDAR scan.

Think of a person using different tools to cook a meal. If a thermometer and an oven timer are not set to the same time, the dish may not come out correctly. Aligning tool readings is like calibrating sensors to ensure they work together accurately.

🔉 Noise Modeling All sensors exhibit noise — random variations that obscure true signals. Modeling noise helps filter and smooth data for better interpretation. Noise Sources: ● Electrical interference ● Environmental factors (lighting, temperature) ● Sensor wear and drift

Noise in sensors refers to the random fluctuations or errors in data collected, which can obscure the actual signals intended for processing. Various factors contribute to this noise:

• Electrical interference can arise from nearby electronic devices.
• Environmental factors like changing light or temperature can affect readings, especially in sensors like cameras.
• As sensors age, they may experience wear and drift, leading to inaccuracies over time. Understanding these noise sources is vital for effective data interpretation.

Imagine trying to listen to your favorite song while standing next to a busy road. The sound of cars passing can make it hard to hear the music clearly. In our case, the cars are like noise in sensor data, and to enjoy the song, similar to filtering noise, you would need to reduce the sound from the road.

Common Models: ● Gaussian noise: Assumes error distribution follows a normal distribution. ● Bias and drift: Long-term deviation of sensor readings.

To deal with noise, certain models are commonly used:

• Gaussian noise is a statistical model where the errors are assumed to be normally distributed around a mean value, which helps in making predictions about likely readings.
• Bias and drift refer to the gradual changes in sensor readings over time, resulting in systematic errors that accumulate, leading to inaccuracies if not corrected. Understanding these models helps in the implementation of better filtering and smoothing algorithms.

Think of Gaussian noise like a crowd where most people gather around the center, but some are much further away. Bias and drift can be compared to a pot that slowly boils over if left unattended; you must adjust it periodically to prevent it from spilling over continuously.

--