Jamming Attack Case Studies

Experimental Setup

For these case studies, the attacker and victim radar were placed 15m away from each other. At the start of the case study, a target vehicle begins to drive away from the victim radar at a velocity of roughly 20mph. (~9m/s).

When reviewing the following results, we make the following notes:

The minimum detection range of the victim radar is 24.35 m*.
Detections corresponding to stationary vehicles and objects are not shown in the plots below**.

Jamming Attack Case Studies

Attacking victims employing parameter randomization

mmRAD is designed to launch attacks against victims with a consistent chirp slope, chirp period, and frame rate. However, it is possible that some victims will employ various randomization techniques in an attempt to thwart our attacks. One of the easiest and most common defenses is to randomize key radar parameters like the frame start times, chirp periods, and chirp slopes. While this is effective at protecting against previous attacks, our novel mmRAD platform is capable of detecting when a victim radar is employing such a defense. Since our primary attacks would not be effective against such a defense, our architecture automatically switches to an optimal jamming attack that seeks to reduce a victim radar's probability of detecting any object in general. While not as powerful or targeted as our primary attack, this still can have a significant impact on radar performance.

Without Jamming Attack

Jamming Attack

No Jamming Attack.mp4

Case Study Movie

Without Jamming Attack

In this case study, we demonstrate the effect that randomization can have on standard spoofing attacks. Here, the victim varies the start time of each frame using a small offset with a standard-deviation value of 0.3 microseconds. Meanwhile, the attacker attempts to launch a translation attack by combining a false negative attack centered at (75m, -10 m/s) with a false positive attack at (75m, 12 m/s).

Detected Object Locations

This figure documents the range that the victim radar perceived the target to be for each frame of the case study. Observe how the victim is able to detect the target even after the attack starts. Furthermore, the attacker's false positive attack is significantly skewed in each frame allowing the victim to easily filter out the inserted spoofed points. Finally, the attacker's attacks completely fail to be detected in some frames due to the randomization.

Attack Timeline

The figure above documents the progression of the case study. Here, the top row shows the Range-Doppler response while the bottom row shows the CA-CFAR detections. The first column shows the radar's perception when not under attack while the second and third columns represent the radar's perception when under attack. Here we highlight that when the victim randomizes their parameters, a majority of the attacks fail. Observer how only three of the attack frames resulted in successful attacks

Jamming Attack.mp4

Attack Movie

Jamming Attack

In this case study, we demonstrate how our attacker utilizes optimized jamming attacks when it detects that a victim is randomizing its parameters. As before, the victim varies the start time of each frame using a small offset with a standard-deviation value of 0.3 microseconds. Meanwhile, the attacker attempts to launch a translation attack by combining a false negative attack centered at (75m, -10 m/s) with a false positive attack at (75m, 12 m/s). In this case though, the attacker detects that the victim is randomizing its parameters and immediately switches to a jamming attack.

Detected Object Locations

This figure documents the range that the victim radar perceived the target to be for each frame of the case study. Observe how the victim fails to detect the vehicle once the attack starts even though it is also randomizing its parameters.

Attack Timeline

The figure above documents the progression of the case study. Here, the top row shows the Range-Doppler response while the bottom row shows the CA-CFAR detections. The first column shows the radar's perception when not under attack while the second and third columns represent the radar's perception when under attack. When no attack is present, we clearly see the vehicle moving away from the radar in the Range-Doppler response where it is also detected using the CFAR. However, when the attack is engaged, observe how the attack has uniformly added added clutter to the victim's Range-Doppler response; the result is that the victim fails to detect the bus when the jamming attack is present.

Footnotes:

*This is a result of the CA-CFAR detector configuration (see Sec.VII in the paper for a more detailed explanation) and hardware limitations. Thus, it is possible that some targets at close range will not be detected until they move further away.

**Stationary vehicles are significantly more difficult to detect as they often get lost in the clutter from the surrounding environment. Additionally, these detections are not shown in order to make it easier to observe the behavior of targets of interest.

Page updated

Google Sites

Report abuse