How to use Artificial Intelligence and Machine Learning to catch drug traffickers

03 Sep 2018

  • Analytics
  • Technology
  • Data
  • Marketing Technology

In its continued pursuit to protect the nation from both foreign and domestic threats, Australia’s national security sector is being challenged to drive greater value from the Internet of Things (IoT) as well as new technology advancements such as Artificial Intelligence (AI) and Machine Learning (ML).

Simultaneously, there is a need to connect the sector’s human capital with these powerful capabilities – without impacting their current mission.

In the final instalment of this three-part blog series, Assistant Chief Patrick Stewart shares insight into the United States Border Patrol’s innovative applications of Artificial Intelligence (AI) and Machine Learning (ML).

Q: How is United States Border Patrol (USBP) using GIS technology to unlock the value of IoT?

A: IoT is something that we’ve been doing for years but no one has called it that until just recently. An example of how we are using IoT is the National Intrusion Sensor Infrastructure (NiSI).

This program leverages thousands of IoT devices to detect seismic, magnetic and infrared activity as it occurs, allowing us to detect and track the locations of suspects, agents, and dangerous wildlife.

Some IoT feeds are associated with producing photos or video clips; but some are as simple as weather data – providing us with windspeed and temperature – which is critical as it helps us understand the potential speed of travel when people are on foot. Obviously if it is really hot, someone will travel at a slower speed than if it’s cold.

We use that to determine where we should intercept them to help ensure our agents and the suspects themselves are safe. In the cases of a drug mule trekking through the dessert, we want to apprehend them quickly and efficiently to ensure public safety.

In the case of human trafficking suspects and immigrant family units, we also want to locate and apprehend them quickly to minimise their risk of suffering in treacherous conditions.

Overall, by using GIS to optimise the data collected through IoT devices, we’ve been able to improve our agent dispatch, blueforce tracking and situational awareness processes.

Q: How has your team been using some of the new technology trends we’re seeing in areas such as Artificial Intelligence (AI) and Machine Learning (ML)?

These are the newest areas we have started to push forward in and there’s lots happening. For example, in terms of AI, we have started using GIS to map and analyse IoT sensor information to determine travel patterns, smuggling patterns and examples of narcotic traffic. From there, we can better position these IoT devices to feed an AI computer vision system.

This system automatically detects whether people have weapons, are hauling oversized backpacks or drug bundles, if children are present, or even if there are dangerous animals or endangered species in the area. With this machine learning and AI capability, we have been able to transform a previously labour-intensive task that required us to inspect every visual, into a situation where we can now assess thousands of images a day.

This means we can quickly identify threats and trafficking activity and intercept large narcotics loads.

This is the final interview with Assistant Chief Patrick Stewart in a three-part series. Read the first interview on how GIS is used to secure US borders from transnational criminal organisations; or read the second interview to learn about their interagency operability strategy that enables seamless joint-missions with the FBI.

This article originally appeared on esriaustralia.com.au

Need more info?