A factory installed probe vehicle sensor platform designed to support real-time traffic and road condition monitoring was estimated to cost $40 per vehicle.

A journal article presents a cost evaluation for in-vehicle hardware and software required for an inertial measurement unit (IMU) based traffic and road condition monitoring system.

Made Public Date
02/25/2020
Identifier
2020-SC00444
TwitterLinkedInFacebook

Summary Information

This paper presents a new type of wireless in-vehicle system designed for real-time traffic estimation and road surface monitoring. The platform is mainly designed for probe vehicles but can be installed easily in vehicles equipped with USB chargers. The sensor platform is built around a 32-bit ARM Cortex M4 micro-controller and a LSM9DS0 Inertial Measurement Unit (IMU) module.

Real-time traffic estimation
Vehicle trajectory is determined using an Inertial Measurement Unit sensing module connected to the vehicle at a fixed angle using a USB interface. Then, the module is calibrated to map the coordinates of the sensor to the coordinates of the vehicle. Once the IMU device is installed and its orientation calibrated, the trajectory of the probe vehicle can be estimated through measurements of accelerations, rotation rates, and magnetic field along its path. A GPS device was included for validation purposes.

IMUs do not require any external infrastructure to work and do not receive or transmit data wirelessly. They require an extremely low power to operate, considerably less than GPS or cellphone-based systems. Owing to their much lower complexity than GPS systems, IMUs are less expensive to manufacture than the latter. They do not require an antenna for receiving signals, and are not at risk of losing connectivity with positioning satellites, which frequently happens with GPS systems, particularly if obstructions are present between the receiver and the satellites. Because of their high accuracy (over short time windows), IMUs are good at detecting and classifying the type of congestion encountered (traffic light, stop and go waves, slow and continuous traffic). In addition, such a system offers strong guarantee for the privacy of the participating users when used in conjunction with a short-range wireless sensor network. Those features of IMUs make it more reliable and potentially more accurate than a GPS based positioning system for traffic measurement purposes.

Road surface monitoring
The IMU was also used to detect and monitor road conditions (pavement roughness). The main source of IMU data used for this process was the vehicle’s vertical acceleration characteristics. In addition, a digital microphone was embedded in the platform to monitor rolling road noise as a predictor of pavement conditions.

Computational requirements
Even though the low-cost IMU can generate high frequency and accurate sensing data, it does not generate speed or positioning data directly, which is necessary for most of the traffic-related applications. Thus the system had to support the following computational tasks for real-time traffic and roadway monitoring:

  • Automatic calibration of the IMU including the calculation of a 3x3 rotation matrix based on linear fitting algorithm.
  • Trajectory estimation, including the least square optimization algorithm for speed estimation and the Direction Cosine Matrix (DCM) filter for attitude estimation.
  • Road surface condition monitoring based on IMU data, including a linear regression algorithm for monitoring and predicting road surface conditions.

Communications
Bluetooth protocol was used to connect wireless components within the vehicle and network nodes at the roadside.

Software
Different sensors are read at different time intervals. Data are stored at a fixed, regular sample rate. A sampling rate of 10 Hz was achieved when running all trajectory estimation algorithms which was assumed to be sufficient given the time scales associated with driving.

Platform cost evaluation
The table below excerpted from the source report shows the cost of the major components (excluding manufacturing costs) in three different in-vehicle platforms All required sensing, communication, and storage modules. With economies-of-scale and significant levels of market penetration, the proposed probe vehicle system was estimated to cost approximately $40 per vehicle.

 

 

Version
Item
Quantity
Price
Breakdown Price
Remarks
First generation



 
STM32F407
1
$11.05
$7.18@1000 Micro-controller
SH-HC-06
1
$8.99
$5.46@500 Bluetooth Transceiver
Beitian BS-280
1
$12.19
$10.79@100 GPS
XBP24CZ7UIT-004
1
$10.40
$8.53@500 XBEE Transceiver
GY-85
1
$8.45
$5.82@300 IMU Sensor
Total
1
$50.72
$37.38@1000  
Second generation


 
STM32F407
1
$11.05
$7.18@1000 Micro-controller
SH-HC-06
1
$8.99
$5.46@500 Bluetooth Transceiver
XBP24CZ7UIT-004
1
$10.40
$8.53@500 XBEE Transceiver
GY-85
1
$8.45
$5.82@300 IMU Sensor
Total
1
$38.89
$26.99@1000  
Third generation




 
STM32F407
1
$11.05
$7.18@1000 Micro-controller
SH-HC-06
1
$8.99
$5.46@500 Bluetooth Transceiver
SIM800C
1
$8.75
$5@500 GPRS Transceiver
Beitian BS-280
1
$12.19
$10.79@100 GPS
LSM9DS0
1
$7.11
$3.83@3000 IMU Sensor
ADMP401
1
$6.99
$4.12@500 Microphone
Total
1
$55.08
$36.39@3000  

 

 

System Cost

Probe Vehicle IMU: $40 per vehicle

System Cost Subsystem