Sensor continuous path identification algorithm - Database & Sql Blog Articles

Single chip microcomputer STM32L151CCU6

Photocoupler

For infrared photoelectric sensors, the receiving tube detects the ground's reflected infrared light and its voltage varies depending on the road surface color—white or black. When the sensor faces a white surface, it generates a higher voltage, while a black line results in a lower voltage. This principle allows for a simple path recognition algorithm: the sensor voltage is read via an I/O port, and the logic level determines whether the sensor is over the marking line. This method helps identify the car’s position relative to the path.

Although this discrete approach is straightforward and requires minimal hardware and processing power, it has limitations. It only provides data from individual sensors, leaving blind spots between them. In races with limited sensors, this leads to reduced accuracy and poor performance when navigating complex paths.

Even with more sensors, the discrete method still suffers from inherent issues. Since it relies on binary (on/off) information, it can cause abrupt changes in steering and speed control, leading to overshooting, oscillation, and delayed responses. This is especially problematic at high speeds, where quick decisions are crucial for stability and control.

To address these challenges, researchers have explored advanced algorithms that provide continuous path information. One promising approach involves using finite sensor arrays to estimate the distance between the vehicle and the path line, offering real-time, smooth feedback. This method improves accuracy by leveraging sensor characteristics and calibration data.

Photoelectric Sensor Characteristics

The behavior of infrared sensors isn't just about detecting white or black surfaces. The voltage output correlates with the distance from the path line. The closer the sensor is to the black line, the lower the voltage, and vice versa. By understanding this relationship, we can determine the exact position of the vehicle relative to the path line, rather than just knowing if it's on or off the line.

Continuous Path Recognition Algorithm

Before implementing any algorithm, it's essential to calibrate the sensors. During pre-calibration, the car is parked, and the microcontroller records the maximum and minimum voltages as the sensors sweep across both white and black areas. These values are used to normalize the sensor readings during actual operation.

In each decision cycle, the sensor voltage is converted to digital and normalized. Then, valid sensors are identified, and their offset distances are calculated based on the pre-defined sensor curve. Averaging the results from multiple sensors enhances accuracy and reduces noise from individual readings.

This method provides continuous path deviation information, eliminating the gaps between sensors. Whether the car is directly above the line or slightly off, the system can detect and respond to changes in real time, improving overall control performance.

Challenges and Future Directions

Compared to traditional discrete methods, the continuous approach offers better precision and smoother control. It can work effectively regardless of the number of sensors, making it ideal for high-speed applications. However, successful implementation requires careful hardware design.

Choosing the right sensors based on track width and mounting height is crucial. Additionally, ensuring consistent sensor behavior across all units helps maintain algorithm reliability. While perfect uniformity is hard to achieve, proper component selection and calibration can significantly improve performance.