The Israeli start-up clearing minefields with artificial intelligence

The Israeli start-up clearing minefields with artificial intelligence

A sign marks a landmine field, left from a past conflict, near Israel’s Golan Heights village of Merom Golan August 20, 2006. Syria, a key Hizbollah ally, wants the Lebanon war to lead to a comprehensive peace settlement that addresses what Damascus regards as the root of instability — Israeli occu.

Despite record global investment in landmine clearance in 2017, exceeding $770 million worldwide, nearly 2,800 people were killed and thousands more injured by mines and other explosive remnants of war (ERW). Civilians, half of them children, accounted for 87% of all casualties.

Over half of 62 countries known to have mine contamination are committed to the Mine Ban Treaty, also known as the Ottawa Treaty, but only four are on track to meet their clearance deadlines. At the current rate of mine clearance, some estimates suggest that it will take over 200 years to clear the world, at a cost of over $100 billion.

While the vast majority of mine clearance actions rely on dangerous manual labor, a start-up founded by two former officers from the IDF Combat Engineering Corps’ elite unit, Yahalom, offers a new approach to mine clearance based on artificial intelligence and autonomous robots.

Established by Itzik Malka and Nir Cohen, 4M Analytics provides a “minefields location intelligence solution” which the company says can enable high-paced mine clearance at a lower cost and with lower risk than existing solutions.

The company is based at Kinneret College and Amazon’s recently-opened AWS Builders Space in Tel Aviv, and boasts staff with expertise from military and civilian mine clearance operations, in addition to data scientists and software engineers.

By integrating machine learning algorithms with a series of old and dynamic new data sets gathered by remote sensing, 4M Analytics aims to precisely identify and reduce suspected hazardous areas. During a two-year research period focusing on Israeli minefields, they found that approximately 80% of each field is actually clear from mines. “Our technology has been validated on three different projects here in Israel, where we found 100% correlation between the hazardous areas that we identified and what we ultimately found,” 4M Analytics chief product officer Raz Ezra told The Jerusalem Post .

“We checked the entire minefield areas as demanded by the Ministry of Defense, but only found mines in the same areas that we defined as hazardous. Using the system at Gofra Beach, near the Sea of Galilee, we also found six landmines beyond the minefield fence in an area open to civilians since 1980.”

The company’s solution relies on a five-stage process of data collection, data fusion, data analysis, a prediction algorithm and technical surveys implemented by autonomous robots.

Data is gathered from historical maps and plans, prior to the estimated date of mine plantation and up to the present day. Dynamic data is collected by remote sensing based on multifaceted technological sensors including LIDAR, hyper-spectral, ground-penetrating radar, electromagnet, ultrasound and thermal imaging.

While the process might at first sound costly, Ezra states that the full package of mine clearance will cost $1.76 per square-meter, compared to usual clearance costs ranging between $4 and $8 per square-meter worldwide, and as much as $13 in Israel.

“Our case study has been the Golan Heights, where there are more than 100 square-kilometers of minefields. If we continue at the current clearance pace of one square-kilometer per year, it will take 100 years to clear the Golan Heights,” said Ezra.

“The cost of the clearance exceeds $1 billion, amounting to 35 shekels per square-meter. We can assure them with our technology to reduce clearance time from 100 years to 10 years, reduce cost from $1 billion to $60 million, and to minimize the 100 square-kilometers to under 20 square-kilometers of hazardous territory.”

To date, the company has completed successful pilots at Gofra Beach and in the West Bank local council of Karnei Shomron, where 64% of the minefield was deemed non-hazardous and all 1,765 existing mines were not functional.

4M Analytics is currently focusing on securing business-to-government deals, including with ministries of defense and United Nations agencies. The company hopes to soon secure pilot projects in Ukraine, Turkey, Nigeria and Cyprus.

The company has also held productive conversations with Paul Heslop, the head of program planning and management at the United Nations Mine Action Service (UNMAS). Together with the global body, the company will execute a pilot project in 18 suspected minefields in Cyprus.

“We are really focused on minefield clearance,” said Ezra, “but we understand that after we have the full proof of concept (PoC) on this specific domain of minefields, we can enter different markets too.”

Future applications touted by Ezra include using the company’s sensors and predictive algorithms to assist infrastructure developers, agriculture, and oil and gas companies to identify suitable and safe locations for construction and operations. Artificial intelligence


What do you think?

48 points
Upvote Downvote

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Deep Fakes' Greatest Threat Is Surveillance Video

Deep Fakes’ Greatest Threat Is Surveillance Video

Artificial Intelligence firm Capacity raises $13.2 million

Artificial Intelligence firm Capacity raises $13.2 million