Women Safety using Unmanned Aerial Vehicles

Comments · 286 Views

With the burgeoning of crimes against women, women security has become a matter of concern. In this world of advanced technology and smart electronics it is required to have an advanced women security system to provide the much necessary safety measures in public places as well as when traveling alone. In this paper, we present one such advanced security system. Currently, there are many women safety apps which take videos, photos, share location, send alert messages for safety concerns. All these applications are only helpful for tracking the location of a person but not in predicting any threat. Sometimes, it might be too late for help to reach the place of threat, if there is any. So, In this article, I would like to share my idea on predicting the threats using a drone surveillance system by monitoring the surroundings of the person and preventing the attack if possible. This is the age of Artificial Intelligence and Machine learning and we can use that to our advantage to make these predictions.

INTRODUCTION:      

In today’s world, safety of humans has become a major concern especially for women. Technology is being used a lot to provide this safety, but all the applications that are currently in the market are mainly focused on location tracking and alerting a person from the list of emergency contacts regarding the status of safety of the victim. With humongous growth in Artificial Intelligence and Machine Learning , it is possible to prevent some of the attacks rather than just knowing about them and locating them as it may be too late to reach the place of incident. So it is better if we predict any such incident well before so that we will have a chance to act. Unmanned Aerial Vehicles(UAV) such as drones can be used to achieve this.

Currently, Aerial Surveillance systems [8] uses drones in areas such as civil and military purposes to prevent people from being harmed and also monitor enemy movements, Industrial surveillance to monitor extreme machines, agriculture for remote sensing, crop product management like application of pesticides and increasing farming efficiency but did not enter personal scope like safety purposes of an individual. We can extend this concept of Aerial Surveillance System to prevent physical attacks using a drone companion. Currently, many of the drones in the market are large in size and can be easily noticeable. It would be better if the attacker cannot easily identify that there is a drone observing the person because the attacker might try to destroy the drone before choosing to attack the person. In order to avoid this, we can consider some of the proposed miniature drone concepts that are being currently developed [4]. We propose that once these miniature drones are available for production use, we embed an eSim into these drones so that they have wireless communication abilities without affecting the form factor. We call this an eSim Nano-drone. These drones can also be cost effective as compared to the drones currently in the market. These types of drones can be used in crowded places to monitor incidents without disclosing its presence and in general we can use it for surveillance at places where the drone's presence should not be known. These drones can be used for general security surveillance at organisations instead of manpower as these are affordable compared to their wages. There is also some work being done where a drone-assisted multi-hop device-to-device (D2D) communication scheme as a means to extend the network coverage over regions where it is difficult to deploy a land-based relay[10].           

The potential of these miniature drones is infinite and although there are many more applications that are possible with nano-drones, we are narrowing our scope to ‘women- safety’ using UAV (nano-drone) in this paper. We connect drones with individual’s smartphones so that the server can send alerts regarding any threat to the individual. If all the technologies we use in this idea come to existence in future, many women can be saved from physical attacks, harassment. This also helps companies and organisations to save expenditure spent on resources when providing physical companions to women working late nights by replacing physical companions with drone companions.

EXISTING SOLUTIONS AND THEIR DRAWBACKS:

All the present applications like Smart 24x7, bSafe, iGoSafely, Trakie, CitizenCop etc., [1] are built on the idea   of sharing live location of the person, recording  videos,  taking pictures, sending alert messages to circle, contacting emergency number and cops. These are helpful to assure whether a person reached the destination safely or not. Some  cases include families tracking their children going to school, tracking women who work late hours, safety of people travelling in unsecured paths and senior citizens in distress. These apps are able to provide evidence of the incident that occurred by recording through camera.

There are  apps  like  Safetrek  [2]  which  allows  you  to press a button when unsafe and later press a pin to confirm your safety, if it is not done your location is automatically shared to local cops. Existing UAVs might be helpful up to certain extent, by selecting an UAV which is suitable to the problem.

DRAWBACKS: All the women safety applications or services cannot predict an occurrence of the attack or help the victim in defending themselves. It can only send the details of location or criminal after committing a crime.Some applications like Smart 24x7 is supported only by certain police stations, which cannot be helpful to women of other regions.

ARCHITECTURE OF IDEA:

The drone companion system consists of a miniature flying machine or drone as the primary component. This flying machine is capable of autonomous flight, streaming or recording video and also to communicate with  the  cloud  server  and the smartphone. The camera attached to the drone will live stream the video of the surroundings of the person to the cloud server. The server uses deep learning techniques to predict any threat/attacks if any and then sends that information to the smartphone.

                                             Fig.1. System Architecture

 

  1. Components

From the above Figure.1, System Architecture, Components required are:

  1. eSim Drone: The miniature eSim drone is the primary component in this architecture. It consists of mounted cameras, motor, rotary rings, GPS receiver, antenna and a computer chip. We can see all the components in the below image.
  2. Cloud Server: Since the drone is not powerful enough to perform the video analysis, this data has to be sent to the cloud  server. Data can be sent to the cloud in many ways. In our model, the drone is embedded with an eSim and the data can be sent using a high speed 5g network.
  3. 5G network: We need 5G network to be deployed on a wide scale so that we can get away with not having a WiFi adapter on board which affects the form factor of our drone.
  4. SmartPhone: The streamed video is analysed  at  the  server and if any threat/attack is detected, it will send that information to the smartphone of the user so that the user  can be aware of the situation well in advance to take the necessary action.

WORKING OF IDEA:

Embedded Sim :

This concept is already implemented in mobiles such as Google Pixel-2, iPhone XS. These devices have a special em- bedded chip which allows users to select the network without physical sims. Advantage of these sims is that the device does not depend on a network’s signals at a particular location. If one network has low signal, this e-sim automatically switches to the other network with high signals. This concept is now being used to build eSim drones. By 2022, 13 percent of the drones are going to be equipped with eSim [3].

In simple terms, eSim can be referred to as integrity of tradi- tional sim cards using secure protocols.Data transfer is done securely using eSim protocols.But the challenge is to combine business logic of various distribution channels into a single technical solution for eSim. However, after certain practical work 2 solutions (consumer solution and M2M solution) came into force that suits different channels. Consumer solution requires the end user to choose the operator supplying connectivity whereas an M2M solution is formulated for business to business customers[7]. The only way this can be achieved is to incorporate GSMA membership in the devices used. This modeling can be depicted from figure 3 and figure 4.

                                       Fig.2. Architecture of Drone (DragonflEye) [4]

 

 Prediction steps

  Prediction is done in a sequence of steps:

  1. It observes the people in the surroundings up to 50 meters for any suspicious behavior.
  2. It short lists people based on body behavior and movements towards the person and also retinal movements
  3. It further short lists based on the distance travelled by them in our direction and closeness
  4. Observes if there are any instruments with them that can be dangerous
  5. Raises an outward warning sirens as a defence mechanism to scare the intruder when approached to the set boundary of victim
  6.   It should be able to detect eve-teasing based on the actions of the woman and the person doing it i.e., A man tries to attack a woman with words for which the woman does not respond. In such cases, it is categorised as teasing.

Along with predicting any danger it also does other activities like:

1.Drones communicate among them through eSim network and informs us about any security issue at some place or incident occurred to prevent us from going  that  way 

2.Sharing global information to authorities for public use 

3.Provides Light when walking in roads which are dark

Communication Process

Drone companions are connected to a smart phone through  e-sim. It informs the person about any potential threat by developing an efficient algorithm for the above mentioned steps and suggests a solution to prevent it from happening. During emergency it connects to the emergency contacts in the smartphone and shares location to them along with alert messages. It can also decide whether to  inform  the  local  cops based on the intensity of the situation. These drone companions are nano sized such as Dragonfly [4].

                                              Fig.3. eSim Configuration [7]

 

                                                   Fig.4. eSim Selection [7]

Coming to identifying threat/attack from the video, currently we have some proposed solutions like Detecting Robbery and Violent Scenarios [6]. The method first extracts the motion  region from the video and denotes this region with a rectangle. Then, the method calculates the optical flow and energy of   the rectangular region. The method takes the length and width of the rectangle, the energy,  and the orientation variance of  the motion region which is denoted by the same rectangle as features to distinguish the video where violent segments occur [6]. So we can use these existing solutions and also these methods can be deployed dynamically on the cloud as better methods are invented in the future and our current architecture is not affected by this.

LIMITATIONS OF THIS IDEA:

Drone requires a lot of energy to function  continuously. The charging should be done in very negligible time so that    it will not deviate from its duty.  It may not work efficiently if the attacker hides behind hoodie or is completely disclosed. While this UAV drone technology is widely spreading in alL sectors, there is also wide discussion about the concerns when using UAVs as it is affecting security of nuclear facilities like kinetic attack, electronic attack, reconnaissance, smuggling and distraction [9].

FEASIBILITY ANALYSIS:

The proposed idea cannot be implemented currently due to  technological  limitations  such  as  availability  of  e-sim  in production, wide adoption of high speed 5G, battery limitations of the drone and need of charging pads. But these technologies are currently undergoing a lot of transformations and are subjected to change and improve in the next 5 years. These limitations are further delineated as below:

1.eSim Drone: 13 percent drones are subjected to be E-sim drones by 2022 according to research reports [3]. So, by 2025, 70-80 percent of drones need to be equipped with eSim for this idea to work effectively. 

2.Drone Charging: As the drone is very small in size, its battery has to be high energy density. If not, Charge stations availabil- ity should be very frequent with charge pads so that drones  can charge immediately.

3. Privacy: If someone hacks into a drone, our information can be easily  accessed.  They  can  also  divert  us  into  danger  by operating drones according to their advantage. Also, the company manufacturing the drone can easily look through our data. Therefore, there has to be a strict privacy policy between manufacturers and users.

4. Weather conditions: It should operate under extreme weather conditions, different light spectrum like infrared at nights, rain.

5. Physical Strength: There might be a case where the attacker  identifies the drone and tries to destroy it before attacking. In such cases, the drone should be able to resist the attack and protect itself by taking reflex decisions sensing the sudden attack.

REFERENCES:

  1. http://www.businessworld.in/article/10-Safety-Apps-For-Women/12-06-2018-151793/ 
  2. https://rocketit.com/safety-apps/
  3. https://www.counterpointresearch.com/cellular-connected-drones-will-iphone-moment-drone-industry/
  4. https://steemit.com/science/@ellisburgin1/dragonfly-drone-concept
  5. https://www.smithsonianmag.com/innovation/turning-dragonflies-drones-180962097/
  6. Y. Xu and J. Wen, ”Detecting Robbery and Violent Scenarios,” 2013 Second International Conference on Robot, Vision and Signal Process- ing, Kitakyushu, 2013, pp. 25-30. doi: 10.1109/RVSP.2013.14 
  7. https://www.gsma.com/esim/wp-content/uploads/2018/12/esim- whitepaper.pdf
  8. https://www.gsma.com/esim/wp-content/uploads/2018/12/esim- whitepaper.pdfZ. Zaheer, A. Usmani, E. Khan and M. A. Qadeer, ”Aerial surveillance system using UAV,” 2016 Thirteenth International Conference on Wireless and Optical Communications Networks (WOCN), Hyderabad, 2016, pp. 1-7. doi: 10.1109/WOCN.2016.7759885
  9. Solodov, Alexander, Williams, Adam, Al Hanaei, Sara, and Goddard, Braden. Analyzing the threat of unmanned aerial vehicles (UAV) to nuclear facilities. United States: N. p., 2017. Web. doi:10.1057/s41284- 017-0102-5 
  10. X. Li, D. Guo, H. Yin and G. Wei, ”Drone-assisted public safety wireless broadband network,” 2015 IEEE Wireless Communications and Networking Conference Workshops (WCNCW), New Orleans, LA, 2015, pp. 323-328. doi: 10.1109/WCNCW.2015.7122575

 

Comments