GPS position simulator for Veins

70 views
Skip to first unread message

jblesa

unread,
Jun 20, 2017, 12:17:16 PM6/20/17
to OMNeT++ Users
Hello,

I'm starting to work with Veins simulator. I would like to receive the cars position in a RSU. I've already know how to do that, but the position is always exact. 

  Veins::TraCICommandInterface* traci = mobility->getCommandInterface();
    std
::pair<double, double> currLonLat = traci->getLonLat(c);

¿Is there any GPS signal simulator that introduces real measurements with an error?

Best regards,

Christoph Sommer

unread,
Jun 22, 2017, 7:06:36 AM6/22/17
to omn...@googlegroups.com
jblesa wrote:
> Is there any GPS signal simulator that introduces real measurements with an error?

None I'm aware of. Are you sure what you want is a GPS error model, though?

Cars are unlikely to wirelessly transmit the raw position they
calculated using a GPS receiver as it's very noisy (see, for example,
DOI 10.1109/PLANS.1994.303322 which discusses the many random processes
involved).

Even for an application as simple as a satellite navigation system
(SatNav), raw samples are filtered and smoothed over time, then
correlated with other sensors such as those monitoring the car's
physical movement to be able to also work in a tunnel, etc.

Advanced Driver Assistance Systems (ADAS) further post-process position
information using computer vision systems for precise lane-keeping etc
etc -- and with autonomous cars just around the corner the amount of
sensors that can serve as input for position information increases even
more. Ultimately, you can thus arrive at position information that is
far more precise than what GPS alone can do.

Best,

Christoph


--
Dr. Christoph Sommer
Distributed Embedded Systems (CCS Labs)
Heinz Nixdorf Institute, Paderborn University, Germany
http://www.ccs-labs.org/~sommer/

Pedro Libório

unread,
Jul 13, 2017, 3:21:58 PM7/13/17
to OMNeT++ Users
If you want one synthetic GPS noise, you can simulate it as a Gaussian normal distribution with a specific mean and standard deviation.
As you have you specific noise model, you can add it above the real position.

Cheers

Pedro.
Reply all
Reply to author
Forward
0 new messages