Sensors are devices capable of detecting a specified measurand and converting an input non-electrical signal into an electrical output signal. Consequently, sensor experiments primarily focus on verifying the relationship between the input and the resulting output; as a result, the majority of experiments in sensor education are of a confirmatory nature. Such experiments provide an intuitive demonstration of a sensor's operating principles and effectively reinforce theoretical instruction. Examples include force measurement using resistive strain-gauge sensors, temperature detection with thermocouples, and displacement measurement using capacitive sensors.However, relying solely on confirmatory experiments is far from sufficient for cultivating students' hands-on skills and creativity; we must therefore leverage existing experimental platforms to develop more design-oriented experiments. The CSY-998 sensor system experimental unit serves as a suitable example to illustrate this. This unit enables the execution of an experiment titled "Characteristics of Hall Sensors—DC Excitation," in which a Hall element is mounted on the unit's vibrating disk and two semicircular permanent magnets are fixed to the top plate, together forming a Hall sensor. The operating principle is illustrated in Figure 1.

  Based on teaching practice in the course "Sensor Principles and Applications," this paper analyzes and explores the course's theoretical and experimental content, as well as its teaching methods and tools, and outlines the reforms implemented. Expanding the teaching content has enriched classroom materials and closely linked the curriculum to real-world sensor applications; this has broadened students' horizons while enhancing their interest and enthusiasm for learning. Furthermore, the introduction of design-oriented experiments has helped foster student initiative and creativity. These represent beneficial initiatives in the teaching of sensor technology.To address the critical issue of drowsy driving, a system combining infrared cameras and pulsed LEDs can be used to locate and monitor the driver's pupils; through image processing and feature extraction, the system analyzes pupil constriction or closure to detect signs of fatigue. Simultaneously, a steering angle sensor tracks changes in the vehicle's heading to determine if it is drifting out of its lane. Additionally, a front-mounted millimeter-wave radar provides real-time data on the distance to the vehicle ahead or other obstacles; if the distance crosses a specific threshold, the system issues visual and audible alerts and—should the driver fail to brake in time—initiates emergency braking to prevent a collision.Traffic flow information detection is a crucial component of intelligent transportation systems. It involves fusing data from various sensors—such as traffic volume, vehicle speed, lane occupancy, traffic density, and queue length—and transmitting this real-time information to a control center, where it is analyzed to issue appropriate commands for intelligent traffic control. These parameters are primarily obtained using inductive loop detectors installed in the road surface. The detection principle relies on changes in the loop's inductance: when a vehicle passes over the coil, the detector is triggered to output a signal, which weakens as the vehicle moves away. Consequently, the passage of a vehicle generates a waveform signal; by counting these waveforms, the traffic volume passing over the coil within a specific period can be determined.Two induction loops are installed in the road surface at intervals of 3 to 5 meters; a pulse count begins when a vehicle triggers loop A and ends when it triggers loop B, thereby determining the time duration—measured in pulse counts—required for the vehicle to traverse this distance, which allows for the calculation of vehicle speed. During the multi-information fusion stage, the vehicle speed is assumed to follow a normal distribution, and a weighted average data fusion method is employed to integrate the speed parameters. In the traffic flow detection phase, the mathematical model for traffic flow distribution is determined based on the fused speed data: if the speed is below 24.5 m/s (indicating traffic congestion), the traffic flow follows a binomial distribution; otherwise, it follows a Poisson distribution. Finally, regarding queue length prediction, the estimated traffic flow and the queue length from the previous time step serve as parameters for a Kalman filter prediction model to forecast the queue length at the current time step.The car anti-theft alarm system operates as follows: when the owner activates the system using RFID (Radio Frequency Identification) technology, the entire system becomes operational. Strain sensors installed at the doors detect opening and closing; glass-break detectors monitor for intentional damage to the windows; pyroelectric sensors detect unauthorized entry into the vehicle; and Hall sensors monitor for vehicle vibrations. These sensors work in concert, transmitting data in real-time to a central processor for analysis to determine the vehicle's security status. If an intrusion is detected, the alarm module notifies the owner by sending an alert via GPRS to their mobile phone, while the execution module triggers flashing lights or the horn to attract the attention of bystanders.