IEEE Consumer Electronics Magazine - March/April 2021 - 25

Figure 8. Performance on real-drive cycle in
Colorado for best solution generated by VESPA and
the baseline configuration for the 2019 Blazer.

using a radar coupled with camera in zone B for
LKA reduces the number of false positives during detections. In Figure 7(a), radars 3,4 and cameras 2,3 placed in zones D and E, respectively,
were sufficient for improving performance of
ACC and FCW by reducing the number of false
positive object detections. The combined optimization of orientation and location with VESPA
resulted in a sensor configuration that maximized performance for each feature.
Our last experiment involved testing the best
sensor configuration from our VESPA framework
and the baseline configuration for the 2019
Blazer on data from a real-world drive cycle over
one hour in Colorado. We focus only on assessing performance for the ACC and FCW features.
Figure 8 shows an image from the real-drive
cycle with data collected by the vehicle from the
radar and camera sensors on it. The figure also
shows a plot of the object occlusion rate (OOR).
The OOR for the baseline configuration was
19.64% (it did not detect 11 out of 56 nonego
vehicles), whereas the VESPA generated best
solution had an OOR of 7.14% (it failed to detect
only four out of 56 nonego vehicles). The results
show the effectiveness of our proposed VESPA
framework in generating higher quality sensor
configurations.

models as shown in our analysis on the Chevrolet Camaro and Blazer vehicles. Further, despite
the sensor locations in the baseline configuration of Figure 7(b) being the most intuitive, the
best configuration is the one generated by
VESPA, showing that even people skilled in the
art of sensor placement may find it challenging
to synthesize a significantly better placement
than that generated by VESPA. We also validated
VESPA with real-drive cycle data to show its
effectiveness for real-world scenarios.

& REFERENCES
1. Taxonomy and definitions for terms related to driving
automation systems for on-road motor vehicles, SAE
International Standard J3016, Jun. 2018
2. R. Spinneker, C. Koch, S.B Park, and J. J Yoon, " Fast
fog detection for camera based advanced driver
assistance systems, " in Proc. IEEE Int. Conf. Intell.
Transp. Syst., 2014, pp. 1369-1374.
3. Robert Bosch GmbH, Bosch MRR Data Sheet, 2017.
[Online]. Available: https://www.bosch-mobilitysolutions.com
4. C. Xu, P. Zhang, H. Wang, Y. Li, and C. Lv, " Ultrasonic
echo waveshape features extraction based on
QPSO-matching pursuit for online wear debris
discrimination, " Mech. Syst. Signal Process., vol. 60/
61, pp. 301-315, 2015.
5. S.E Li et al., " Kalman filter based tracking of moving
objects using linear ultrasonic sensor array for road
vehicles, " Mech. Syst. Signal Process., vol. 98,
pp. 173-189, 2018.
6. M. Jamaluddin et al., " An analysis of sensor
placement for vehicle's blind spot detection and
warning system, " J. Telecommun. Electron.
Comput. Eng., vol. 8, no. 7, pp. 101-106, 2017.
7. T. Kim and T. Park, " Placement optimization of multiple
lidar sensors for autonomous vehicles, " IEEE Trans.
Intell. Transp. Syst., vol. 21, no. 5, pp. 2139-2145,

CONCLUSION
In this article, we propose an automated
framework called VESPA that is capable of generating sensor placement and orientation in modern semiautonomous vehicles. VESPA has the
ability to optimize locations and orientations for
a set of heterogeneous sensors on a given target
vehicle. The framework can be tuned to improve
perception on a desired collection of test cases.
VESPA is also scalable across different vehicle

March/April 2021

May 2019.
8. W. Hudson et al., " Multi-LIDAR placement, calibration,
co-registration, and processing on a Subaru Forester
for off-road autonomous vehicles operations, " Auton.
Syst.: Sensors Process. Secur. Vehicles Infrastructure,
p. 7, 2019.
9. D. Shapiro, " Levelling Up: What is level 2 automated
driving? " 2019. [Online]. Available: https://blogs.nvidia.
com/blog/2019/02/06/what-is-level-2-automateddriving/

https://www.bosch-mobility-solutions.com https://www.bosch-mobility-solutions.com https://blogs.nvidia.com/blog/2019/02/06/what-is-level-2-automated-driving/ https://blogs.nvidia.com/blog/2019/02/06/what-is-level-2-automated-driving/ https://blogs.nvidia.com/blog/2019/02/06/what-is-level-2-automated-driving/

IEEE Consumer Electronics Magazine - March/April 2021

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - March/April 2021