Deploying sensor data for livability and safety

This article was written by: Marijn Biesiot, Tim Jacquemard & Rinie van Est

Ears and eyes everywhere
Examples of digital sensors are the camera and GPS on a smartphone. Sensor data can help the police to do their work better, for example in "smart" detection and effective enforcement.(1) Sensor data also allow citizens and businesses to do detective work or take action to make their own living environment more livable and safer.

Who gets to watch whom?

The frequent use of sensors raises questions. Who is actually allowed to monitor whom? What about trade-offs around security and privacy? In this article, we explore the use of sensors to promote livability and safety (2).

We first discuss what sensors the police use. Then we outline four forms of surveillance with sensors: surveillance, sousveillance, horizontal surveillance and self-surveillance. In doing so, we show the diversity of sensors that police, municipalities, citizens and businesses use to try to make society more livable and safer. We then show how sensor data are used in various core police activities. From this perspective, we identify five notable trends in the use of sensors to promote livability and safety. We conclude with a brief summary of the insights obtained.

Police Sensors
In the late 1990s, sensor data consisted mainly of video footage from surveillance cameras on the street. Currently, police are experimenting with cameras with facial recognition and "anti-sensor sensors" that detect devices that interfere with signals (so-called "jammers"). These innovative technologies are currently used only by specialized police teams, but do show three directions of development. Digital sensors have become more comprehensive, mobile and smart in recent years.

Digital sensors have digitized all our senses: they can also "smell," "taste," and "feel.

Expanded
Traditional imaging cameras and microphones can "see" and "hear. The functions of digital sensors have expanded and in fact digitized all of our senses: they can also "smell," "taste," and "feel. Digital sensors can often perceive more than humans can, for example detecting metal and reading DNA for forensics.(3) Police are also investigating digital sensors that track smartphones via wifi signals ('wifi sniffing').(4)

More mobile
In 1975, the Dutch police first mounted cameras in police cars.(5) Since then, digital sensors have become smaller and smaller. There are even nanosensors, such as the "e-noses" used in the Port of Rotterdam to detect unhealthy air and dangerous gases. As a result, digital sensors are becoming more mobile. They are in smartphones, can be worn on the body (like smart watches and glasses) and controlled remotely. "Bodycams," for example, are video cameras that sit on the uniform. This allows the control room to watch and listen in live with an officer. In 2008, the first bodycams were used by the police in Maastricht.(6) Since then, hundreds are in use, albeit still of an experimental nature.(7) The police also use helicopters and drones that are equipped with various sensors, such as a thermal camera to search for marijuana plantations and an imaging camera that can bring a large area into sharp focus.(8)

Smarter
A camera that can recognize faces is an example of a "smart" camera. That camera not only records video images, but could potentially recognize faces of soccer hooligans on those images, for example. Automatic license plate recognition or Automatic NumberPlate Recognition (ANPR) is another example. This camera is "smart" because it has been given software that can read license plates from cars and compare them with license plates from a database. The police are currently testing smart video cameras that can detect appending or calling motorists.

Citizens and businesses own about 1.5 million security cameras.

Four forms of sensor surveillance
In addition to the police, citizens, businesses and municipalities also use sensors to improve livability and safety. The police own about 500-1000 cameras and municipalities have over 3000 surveillance cameras on the streets. In comparison, citizens and businesses own about 1.5 million security cameras.(9) If necessary, the police would like to use that sensor data. Therefore, the police are asking if citizens and companies would like to register their security cameras. The idea behind the Camera in Picture project is that if the police have an overview of all cameras, they can quickly retrieve all data from the cameras near that crime after a crime. About 200,000 cameras have now been registered with the project. This project is a form of public-private cooperation to make society safer. It also shows that citizens are not only monitored by the government, but also regularly operate the camera themselves. From the citizen's perspective, there are four forms of surveillance with sensors:

Four forms of sensor surveillance. Source: Rathenau Institute

1.Surveillance
Surveillance is surveillance from "above. It involves authorities monitoring citizens and objects. Street camera surveillance is a well-known example. In this project we look at surveillance by the police, municipalities and companies. Like the police, municipalities and companies also use other types of digital sensors to improve neighborhood safety. More and more municipalities are experimenting with sensors in "smart city projects. For example, municipalities are using sensors for crowd control. In 2017, the municipality of Amsterdam deployed Wi-Fi sensors and smart camera surveillance to measure how crowded the streets were. In doing so, they tried to manage the Christmas crowds. The Eindhoven municipality is working with the police, businesses and knowledge institutions to reduce nuisance and incidents in the Stratumseind nightlife area. There they are experimenting with sound cameras that not only measure noise levels, but can also alert the officer on duty if they detect aggression. In these "smart-city projects," municipalities collect data about citizens in public spaces. The municipalities of Amsterdam and Eindhoven have formulated principles for dealing carefully with this kind of sensor data. Furthermore, government foundation Geonovum has made a preliminary proposal for a set of ground rules for sensors in public space.

Corporate camera surveillance is not limited to office buildings and parking lots. ProRail is installing smart cameras along the tracks to spot copper thieves or people walking along the tracks. Soon all seven hundred NS Safety & Service employees will wear a bodycam, which they can turn on in the event of an unsafe situation. Another form of sensor technology is the security scanners at Schiphol Airport. These use millimeter wave technology to check whether travelers are carrying prohibited items on their bodies. Sometimes employers monitor their employees with digital sensors. The BestDriver app by DHL Express, Rotterdam Municipality and CGI uses sensors and serious gaming to encourage drivers to drive more sustainably and safely.

2. Sousveillance
In sousveillance , citizens monitor authorities. This is surveillance from "below," such as citizens filming police action with their phones. Many "arrest videos" can be found on YouTube. Such videos can play an important role in criminal cases, such as in the 2015 death of Mitch Henriquez after a police arrest.n During the arrest, police used force. Bystanders made videos as the man lay on the ground, and Mitch Henriquez's family released photos taken at the hospital. The man's death led to demonstrations against police brutality in The Hague's Schilderswijk neighborhood. Two police officers were convicted of assault that resulted in Henriquez's death. Video footage and photographs were used in the trial to show exactly what happened.

3. Horizontal surveillance
Citizens who "spy on" each other, such as watching their neighbors with drones, engage in horizontal surveillance. This phenomenon is also known as "little big brothers . Programs such as "Idioten op de Weg" broadcast images of digital cameras on the dashboard of cars. Some 250,000 cars in the Netherlands already have such "dashcams. In addition to security cameras in the home and at the front door, people are also installing sensors that can detect movements and attempted break-ins. The Interpolis Home Watch, for example, consists of a sensor package with smart camera and sensors for doors and windows. When residents are not home and the sensor detects movement at the door, residents automatically receive a notification on their cell phone. If the movement comes from the mailman or a spider web then they don't have to take any action. But if the notification shows it is a potential burglar, then residents can use an app to call their network or a professional security guard for help. Through WhatsApp Neighborhood Prevention groups, neighborhood residents can alert each other to suspicious situations in the neighborhood, for example, by sharing photos of suspicious people.

4. Self-surveillance
Citizens can also deploy devices and applications with digital sensors that help them comply with livability and safety rules. Such as Fairzekering or the ANWB's Safe Driving app. Sensors measure motorists' driving behavior, and people who drive safely get discounts on their car insurance. This should improve road safety. This is a form of self-surveillance. This sensor technology still leaves room for the user to decide what to do with the feedback. If you turn on the ASR Drive Safely app or KPN's Safe Lock, the app ensures that your Whatsapp and other messages do not get through when you are cycling or driving in the car.

The value chain of sensor data
Many of the examples we discuss actually go beyond data collection by digital sensors. They involve digital technologies that process this data and also intervene in the physical world. The so-called "cybernetic loop" helps understand how this works.

Cybernetic loop
The three steps of the cybernetic loop are: data collection, analysis and application (see Figure 2). The cybernetic loop visualizes the entire value chain of digital sensor data. An example: automatic number plate recognition(ANPR). The ANPR camera collects data: it continuously records video images of cars on a highway. The camera is connected to software that analyzes the collected images. The software can identify license plates and compare them with a database in which the license plates are linked to the car's owners. When the algorithm identifies a "hit," several applications are possible. An officer sees the hit and takes action. The 'smart' camera can also take action itself by, for example, automatically sending a fine for a speeding violation.

The three steps of the cybernetic loop broadly correspond to three core activities of police work: witnessing, detection and enforcement:

Cybernetic loop: the value chain of sensor data. Source: Rathenau Institute

Collecting data: witnesses
Digital sensors are a kind of digital witness to situations where livability or safety is under pressure. The information that sensors provide supports and enriches police observations. Citizens, companies and municipalities collect a lot of sensor data that can help increase safety in society. The police must ask them for permission to see this data. This also applies if the police want to view images from registered security cameras in the Camera in Picture project. In addition to sensor data, the police collect information from open sources (such as YouTube and Twitter) and closed sources (such as court data and banking systems). Furthermore, the police ask participants in participation platforms such as Burgernet and Amber Alert to provide information about suspicious or missing persons.

Cybernetic loop: the value chain of sensor data. Source: Rathenau Institute

Analyze: detect
Next, police can analyze the collected sensor data and possibly combine it with other data sources. Detection is about searching the data for patterns and suspects. This analysis can be done by police analysts, but also by artificial intelligence and big data analysis. In Roermond, the police are doing an experiment with scientists and the municipality to counter gangs of itinerant criminals who rob stores, pickpocket and burglarize. The Designer Outlet in Roermond attracts many gangs from Eastern Europe, especially Romania. Police are investigating how digital sensors (such as sound cameras, CCTV, ANPR cameras and Wi-Fi trackers) can help identify and track such gangs earlier. In a data center at Eindhoven University of Technology, the datasets are linked and analyzed. With the help of (manually developed) algorithms, they look for patterns in the collected data, such as whether a car comes from Germany, whether it has a Romanian license plate and how many people are in the car.

Apply: enforce
Based on the analysis, the police proceed with an intervention. Enforcement can also focus on crime prevention. The aim of the experiment in Roermond is to recognize suspicious behavior at an early stage and thus prevent crime. The police collect a lot of data for this purpose, including from many innocent motorists in Roermond. Automated data analysis helps officers determine what is deviant behavior and what is suspicious behavior. The algorithm assigns points to cars based on a list of criteria for suspicious behavior. The more points a car gets driving to the outlet, the more likely it is that police will go after it. To protect privacy, data from non-suspicious cars are deleted after analysis. Practice has yet to show how often the algorithm gets it right.

Five Trends
The cybernetic loop provides insight into how sensor data is used in various core police activities: witnessing, detection and enforcement. This line of thinking can also be applied to other parties that use sensor data to promote livability and safety, such as businesses, municipalities and citizens. Not only do they collect sensor data that may be relevant to the police, but they can also analyze this data themselves and take action on it. We identify five notable trends in the use of sensors to promote livability and safety.

Trend 1: There are more and more police sensors and sensor data
The police are increasingly working with sensors. In the late 1990s, police also used sensors, such as cameras to photograph crime scenes and surveillance cameras on the street. But the number and type of digital sensors available to police has grown tremendously. Police officers on the street are supported by digital sensor technology. This affects the role of the police officer. Officers make decisions partly based on information coming from sensors. In this sense, police actions are informed by data. It can be imagined that in the near future sensor data will take on a more guiding role. In such a case, can you still say that the police officer himself makes the decision?

Trend 2: Police automate some of their core activities with smart sensor technology
A smart sensor can also perform core police activities . By smart sensor, we actually mean an automated robotic system. In fact, the definition of a robot is a "machine that can perceive, think and act" (10).

In that case, the three steps of the cybernetic loop are automated. An ANPR camera can collect images continuously. A police officer still decides what action is needed based on an analysis of these images. A completely automated intervention is also conceivable. For example, the Central Judicial Collection Agency (CJIB) could have ANPR cameras automatically issue fines at route checks. In that case, the ANPR camera system collects the data, analyzes it and takes action without human intervention.

Trend 3: Citizens, businesses and municipalities are collecting more and more sensor data
Over the past 20 years, the use of digital sensors in society has exploded. Citizens and businesses own a thousand times as many security cameras as the police. That's still without all the smartphones and other sensors. In practice, that is very many digital witnesses that can support police work.

Trend 4: Police seek new forms of cooperation to use sensor data from society for livability and safety
In this article we give several examples in which the police cooperate with citizens, companies and municipalities to use sensor data to promote livability and safety. In Roermond, the police collaborate with the municipality, Eindhoven University of Technology and the Public Prosecutor's Office. This is a form of cooperation within the public domain. The police also cooperate with private parties: citizens and companies. The police are looking for new ways to unlock sensor data from private parties, such as in the Camera in Picture project. This is a form of public-private cooperation with citizens and companies. Another example is that the police are encouraging citizens to do their own detective work.

Police are developing new digital platforms that facilitate citizens to conduct their own searches. SamenZoeken is an app that helps citizens search smarter for a missing family member, friend or neighbor. The police officer who came up with SamenZoeken says of the app, "It's a substantially different take on citizen participation. You don't ask citizens to help the police in searches; we help citizens in their search.' If the police take over a search, citizens can easily transfer the information gathered. Another example is Automon, to be launched in 2019.(11) It is a kind of Pokémon Go for stolen cars. It works as follows: ANPR cameras on the street recognize the license plates of stolen cars and automatically send a notification to Automon players in the neighborhood. These start looking for the car. The first to find the stolen car gets a reward. An app has even been announced that gives citizens tips to play detective themselves after a minor incident, such as vandalism or burglary (Sherlock).

Trend 5: Private parties do their own detective work and enforce with sensor data
The use of sensor data by police and municipalities are sensor applications in the public domain. Private parties (citizens and companies) are also deploying sensors and sensor data to make their own living environment more livable and safer. We see that citizens and companies not only collect sensor data to share with the police (witnesses), but also analyze the data themselves (detection) and take action (enforcement). Here, we discuss the examples of a car insurance company anddo-it-yourself policing.

Someone who takes out ANWB Safe Driving Car Insurance gets a sensor system in the car that collects data during all trips and tracks driving style. For example, driving too fast is considered unsafe. The safer a person drives, the more discount he or she gets on car insurance. The terms of use state that ANWB may terminate this insurance if the insured does not drive safely: 'You can tell whether you drive safely by the color codes you receive from us in the feedback. We work with categories that are each linked to a color. (...) If in a year you receive six red messages on the speed section and/or one black message on speed then we may cancel your insurance.' So the auto insurer takes action based on an analysis of sensor data: it terminates the insurance of someone who drives unsafely. There is also bycatch. The ANWB has recovered several stolen cars in recent years via these dongles - a USB stick with a SIM card in it to receive mobile Internet on a computer or laptop - and some parents, through reports of speeding, have caught their children using the car unsolicited.

Citizens cooperating with the police is not new. What is new is that citizens are using sensors not only to provide information to the police, but also to do their own sleuthing and enforcement. This kind of "do-it-yourself-policing" happens on its own initiative. Thief spotter Jaime, for example, has been traveling to Amsterdam for years in his spare time to catch pickpockets and shoplifters. He searches the streets for people acting suspiciously, follows them, calls the Amsterdam police pickpocket team, and waits for the officers to handcuff the crooks. All this he films. Since last year, he has also been putting the videos online. His YouTube channel 'Boevenspotter' has more than fifty thousand followers. The police do not find it a problem that Jaime sometimes goes after suspects himself and takes the stolen items, as shown in a video. 'That is his own responsibility. In principle, citizens are allowed to arrest someone caught in the act.'

Own Judges
Forms of digital citizen participation where citizens not only provide information but also investigate and sometimes enforce themselves raise questions. In addition to success stories, there are also warnings from experts about citizens playing the role of their own judge, getting in the way of the police, getting themselves into dangerous situations through their sleuthing, suffering trauma when they find a body, or invading the privacy of others. Meanwhile, "crook spotter Jaime" makes sure the pickpockets he films no longer appear recognizable in the picture. Imaging suspects recognizable can have implications for criminal prosecution. In early 2013, youths assaulted a man in Eindhoven. The judge gave these youths (the "headbangers")a lower sentence because of video footage. The judge ruled that their privacy had been violated because the video footage was shown on television.

Finally
This exploration of the use of sensors reveals a complex network around "sensor surveillance. Citizens are not only monitored by the police and other institutions(surveillance), but also use the camera themselves in various ways(sousveillance, horizontal surveillance and self-surveillance). The collection of sensor data, its analysis and the action that follows can be carried out by public and private parties and (partially) automatically.

Looking at the entire sensor data value chain, we see five trends in the use of sensors to promote livability and safety:

• There are more and more police sensors and sensor data;

• The police are automating some of their core activities (witnessing, detection and enforcement) with smart sensor technology;

• Citizens, businesses and municipalities are collecting more and more sensor data;

• Police are seeking new forms of collaboration to use sensor data from the community for livability and safety;

• Private parties will do their own sleuthing and enforcement with sensor data.

So we see sensor data being used by public parties (the police and municipalities), private parties (citizens and businesses) and in public-private partnerships. In the next article, we explore in more detail how these different social practices take shape.

These five trends raise questions. Do citizens view police officers with bodycams differently than when NS security personnel wear them? Do citizens care whether they receive a fine from a cop on the street or a smart ANPR camera? Would citizens want to add a feature to their security camera at the request of the police that can recognize faces or recognize people by the way they walk? Do people find it acceptable if pickpockets and burglars are shown recognizable on YouTube? How do citizens view companies that pick up enforcement tasks with sensors?

In this project, we investigate what trade-offs people make when considering these kinds of questions. In which cases do citizens find the use of sensor data desirable (or not), and why? That this can literally get very close is demonstrated by the public discussion about security scanners at Schiphol Airport. When security scanners were introduced a decade ago, there was a fuss about these "naked scanners. A scan of the body showed a person's body and whether that person had hidden prohibited items under his or her clothing. In the process, however, the face and genitals were obscured. Why did some people prefer the scanner and others prefer to be searched the old-fashioned way by a security guard? Was it because of the way people had to "expose" themselves in the scanner or because of other emotions and arguments? More on that in the next article.

(2) The study "Citizen's Perspective on Using Sens Data for Liveability and Safety" focuses on a broad spectrum of liveability and safety. By livabilitywe meanminor offenses, such as throwing garbage on the street. At the other end of the spectrum are more serious crimes that cause a great sense of insecurity, such as street robbery, threats, assault and serious forms of crime such as drug and human trafficking. This study thus stays close to the mission of the police, which at its core focuses on law enforcement and emergency response. Therefore, when choosing to discuss a real-life example, we always ask the question: is this a situation for which you might call the police?

(5) Flight, S. (2016). 'Policing and imaging technology: uses, returns and challenges'. Justice Explorations 42, no.3, pp. 68-94.

(6) Flight, S. (2016). 'Policing and imaging technology: uses, returns and challenges'. Justice Explorations 42, no.3, pp. 68-94.

(7) De Leeuw, P. & I. Nap (2018). Program Sensing. Draft program plan 2018-2019. Draft, version 2.0 February 19, 2018.

(8) Wide Area Motion Imagery (WAMI) is explained on pp. 75-77 of: Police Academy (2018). Knowledge for tomorrow's police. A conference on research at, to and for the police. The Hague: Sdu publishers.

(9) Leenaers, H. (ed.) (2016). The Bosatlas of security. Groningen: Noordhoff Publishers, p.31.

(10) Van Est, R., D. Bunders & I. Korthagen (2017). Rise of robot city politics: The state of affairs in the etherlands. The essay was presented at URBAN AUTOMATION, an international workshop in Sheffield, United Kingdom, on September 4-6, 2017.

(11) Hoorweg, E. et al. (2018). Trust and distrust in digital society.Trends in security 2018. Utrecht: Capgemini.

Source: Ratheneau Institute

This article can also be found in the Internet of Things dossier