Why HIKOB INSIGHTS ?
HIKOB provides real-time information about how the city, its infrastructure, industry, and environment operate to make real-time decisions, learn from reality, and anchor strategies in the real world.
HIKOB’s mission is to make field data available and bring the concept of “time-to-insight” to that of “real-time insight”.
In this state of mind, we are committed to sharing, accurate knowledge and provide the clarifications that we think are necessary to better appreciate the unparalleled level of intelligence that IoT coupled with AI makes us achieve.
In this respect, each month, we will submit an “enlightening focus”.
In this first lighting, HIKOB invites you to perceive the subtleties of vehicle detection and understand how HIKOB leads you to get the best out of it.
What is a wireless magnetometer?
Magnetometry consists of measuring the ambient magnetic field at a given point. The latter is made up of the earth’s magnetic field modified by the presence of surrounding metal masses and / or phenomena disturbing it at this point.
What is a magnetometric measurement?
A wireless sensor is an integrated electronic and software system that makes it possible to acquire and process the signal coming from the sensitive magnetometer component. This one measures the magnetic field at a given point and transmits the data and / or information remotely.
Figure 1 : General architecture of the integrated HIKOB WISECOW sensor
The vehicle detection HIKOB's view
How does a HIKOB WISECOW sensor behave ?
The integrated HIKOB WISECOW sensor regularly measures the local magnetic field in its three dimensions at a specific frequency. When a vehicle passes close to the integrated sensor, it disrupts the local magnetic field. HIKOB WISECOW integrated sensor magnetometer measures this variation, which is then interpreted by the detection algorithm (differentiated according to the type of application: traffic or parking) then communicated over the wireless network.
Figure 2 : Variation of the ambient magnetic field at a given point
What is the function of the detection algorithm?
The measurement of the magnetic field can vary for different reasons:
- The intrinsic properties of the magnetometer sensor:
- The sensor measurement noise,
- The sensor temperature drift.
- The appearance / disappearance of a metal mass near the sensor,
- Changes in the environment, especially electrical (transformer, high voltage line, presence of strong current …).
All the intelligence of the HIKOB detection algorithm lies in its ability to discriminate the variations related to these different causes and to identify the events corresponding to the appearance / disappearance of a metallic mass; then determine in this subset, those generated by a vehicle. In view of the extremely varied vehicle detection needs for parking applications and those related to traffic control / monitoring, HIKOB has chosen to differentiate two algorithmic models.
This choice of design takes form through the implementation of two embedded firmware: the first, for static detection, embedded in the integrated sensors HIKOB WISECOW-P, the second, for dynamic detection, embedded in the integrated sensors HIKOB WISECOW- T.
The power of the HIKOB POLAR STAR version
Option to configure algorithm decision thresholds
To make it easier to adapt systems to specific applications, HIKOB opens access to HIKOB WISECOW integrated sensor firmware sensitivity level settings. Thus, it is possible to influence the detection sensitivity of the system, depending on the application need.
Lowering the threshold amounts would reduce the risk of a “false negative”, corresponding to an undetected vehicle, and favor the risk of “false positives”, corresponding to a detection while the place is empty. On the contrary, increasing the threshold amounts would thus reduce the risk of “false positives” and favor the risk of “false negatives”.
The boundary tests of the HIKOB systems led to significant evolutions of the signal processing model of the firmware dedicated to static detection (HIKOB WISECOW-P).
The initially independent processes of magnetometric detection and calibration monitoring have been merged into a single new algorithm. At the same time, the contrast between the signal and the noise has been increased. The new algorithm incorporating a better response to small variations added to the gain in stability of the calibration follow-up has the effect of excluding the rare losses of state change “false positives” and / or “false negatives”, which have been identified during stress tests.
Figure 3 : Description of the phenomenon “false positives” and “false negatives”
Figure 4 : Situations of “false positive” and “false negative” state change
The auto calibration algorithm compensates for magnetic field drift and the sensor magnetization. When the drift is considered by the algorithm to be too important, a notification on the HIKOB NETPULSE APP (not available with HIKOB NETPULSE) interface suggests a manual recalibration.