Detecting the Presence of Electronic Devices in Smart Homes Using Harmonic Radar Technology

Round 1

Reviewer 1 Report

Thank you all for this work I just have a few questions that will make me understand and evaluate the work better:

• Radar is defined as detection system that uses radio waves to determine the distance, angle, or velocity of objects. Since this paper is concerned with detection what is your justification of using the word radar to describe your system.
• The justification described in 3.2 for using 2.4 Ghz frequency is that

‘Lower frequencies have longer wavelengths and are able to travel much farther, and through more clutter, than higher frequencies’

In a nonlinear system there should be no clutter since you are transmitting f1 and receiving 2f1. In addition, in any nonlinear system the insertion loss at the diode or other nonlinear components is very high you will need the high power provided by lower frequency.

Another disadvantage of using WIfi frequency is that you could easily get interference from wifi signals that will also have harmonic effects. This will cause false detection in your system.

Please justify the use of the 2.4 Ghz frequency.

 

• Also in 3.2

 ‘Radars are generally designed to look for targets kilometers ahead of the transceiver.’

That is not true a lot of automotive radars are designed to work at much lower distances for car park assist and collision detection.

• On page 7 of 17 (according to the copy that I have) the equation 2ftx is written as an equation which I think is wrongly placed.
• Please indicate the sensitivity of your receiver
• Also please indicate the calibration process.
• It is also important for the reader to see the signal received without any object in the same graph with the existence of the device in Figure 6.
• How could you prove that the signal you receive is not simply a reflection of your own harmonics that you send on the object. A metal surface should be placed in the position of the device to compare the reflections.
• Please explain the blue and the red colours in Figure 7
• Figure 11 is very strange for me, from what I understand the authors are increasing the Tx power and measuring the RX power. The linear increase is understandable, what is very strange is the drop you see at about -9 dB.

The justification given is

‘However, for each target, there exists an input power level beyond which  the nonlinearity (e.g., diode junction, transistor input) is maximally activated. In the context of magnetic cores for transformers or RF power amplifiers, this point is known as saturation: no additional field or power may be elicited from the device’

But -9 dB is quite low to saturate the diode or an amplifier especially in a wireless situation. Also the effect of the distance is very strange you have higher RX power from the 35 that the 15. Even after the -9 dBm the 95 cm received power is higher that the 15. Please explain these results.

• Lastly the scope of this paper is to detect the presence of the ‘unwanted’ devices. Please explain how your system will differentiate between the wanted and the unwanted devices that exist in the room.

 

 

 

 

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors have written a very interesting paper on the use of detection of electronic devices by transmission of a RF tone and then detection of harmonics of that tone being generated by non-linear electronics within the propagation environment.  The authors perform experiments with direct connection to the device of interest and then RF transmission.  From these experiments, they show that they are able to detect many different types of electronic devices.

I only have a few suggestions for changes:

1. I would recommend that the Related Work section be moved to inside of the introduction.  Its current place is a little incongruous with the sections before and after it.
2. The harmonics within the transmitters and receivers of the experimental equipment could make the detection method more difficult to make work.  It would be nice to have more information on the level of these harmonics.  I understand that the paper describes the filters to suppress these harmonics but if the harmonics are in the receiver digital sampling hardware, these filters might be of less use.
3. It would be nice to see if RF-shielding on the devices to be detected would make them harder to detect.  Clearly wireless devices would have some frequencies that cannot be shielded but some notes on this would be useful.
4. Some more notes on the geometry of the antennas to the devices on the effectiveness of the detection.  The paper describes the effect of distance but it would be useful to see if the alignment has some effect.  The paper notes that the authors will look at this in future work but it would be nice to see if there were any cases where the detection failed. 
5. The paper indicates that the non-linearity of the devices are what make the detection possible.  There might be some effects of the size of the devices in question with respect to the radio frequency used to detect the devices.  Some notes on this would be useful as well.

Overall, this is a nice paper with some good work which would be of use to many people.  I support its publication with a few notes on the issues above added.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper demonstrates the use of harmonic radars to detect consumer electronics in a home setting and evaluates the performance in both wired and over-the-air transmission scenarios.

I think that it can be published by modifying some of the following.

1) Equation (1) is considered to be an error (reflected field = incident filed x cos wt ?), and why are (2) and (3) different?

2) What is the tx power in Figure 6?

3) In line 364, If the distance to the target is changed, it is unlikely that all sides will be 15cm, and ‘35cm’ in Figure 9 gives confusion. Figure 9 also seems to be an equilateral triangle, so 35cm in the caption may be an error.

4) I think Equation 2 in line 362 should be Equation 4.

5) In line 362, “increasing the area of the target 363 increases the strength of the response (σ in the equation)”. I think this is about conventional RCS and is not related to the nonlinear response. Rather, Is it something to do with the number of nonlinear components or not?

6) If it is electronics that make harmonics, why are the results (order of reflected power, if not absolute) different between Wired and over the air transmission? (Even with the exception of rusty nails)? What is the consistency between the two tests? In Figure 11, Rx power decreases as the distance decreases at the same tx power (-10 dBm), except in the case of 35cm. Can you explain it?

In the paragraph from line 406, you wrote that different nonlinearities or saturation could occur when the Tx power increases. If it is about power, I think, a similar phenomenon should occur when the distance is decreased for the same Tx power. And is it not easy to determine by Figure 12.

In addition, the S band (2GHz) has a wavelength of about 15 cm. So the experimental setup is in the near field (as mentioned in the paper) and the size of the object is similar to or smaller than the wavelength. RCS (linear system response) in this region may fluctuate significantly. It is suspected that there may be harmonics appearing in the receiver (instrument etc.) due to the influence of the TX frequency response(linear response).

It would be better to observe or compare with the response at Tx frequency, in future experiments.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

x