According to Armin Wedler of the German Aerospace Center (DLR), automated planetary exploration methods like those pioneered by recent Mars probes are proving useful in Earth\u2019s disaster areas and conflict zones.<\/p><\/blockquote>\n
Hence Projekt AHEAD \u2013 for Autonomous Humanitarian Emergency Aid Devices. The DLR humanitarian project feeds Earth observation satellite data to remote-controlled vehicles.<\/p>\n
This Martian-inspired application of artificial intelligence (AI), satellite data links, and unmanned aerial vehicles (UAVs) expands the delivery area for relief and assistance, especially during the highly dangerous \u201clast-mile\u201d leg of humanitarian supply chains, explained Wedler during a recent AI for Good\u00a0webinar<\/a>.<\/p>\n As the security risk increases in a given operational area, drones and UAVs can help complete last-mile deliveries of humanitarian aid. Image source:\u00a0Projekt AHEAD<\/a>.<\/p>\n Getting AHEAD with UAVs<\/strong><\/p>\n Equipped with 360-degree cameras and LiDAR technology, Projekt AHEAD vehicles can drive \u2013 and even swim \u2013 autonomously in harsh environments.<\/p>\n A landing pad atop each vehicle enables accompanying drones get a bird\u2019s eye view to protect rescue workers potentially exposed to danger during humanitarian aid deliveries.<\/p>\n \u201cCritical rescue operations can deploy UAVs to overcome shortcomings in infrastructure in inaccessible locations,\u201d pointed out Bilel Jamoussi, chief of standardization study groups at the International Telecommunication Union (ITU).<\/a><\/p><\/blockquote>\n These \u201cmechanical saviours\u201d bring the necessary flexibility and efficiency to disaster areas and conflict zones for surveying damage, identifying survivors, and delivering relief, from food to critical supplies, he added.<\/p>\n DEEP damage assessment<\/strong><\/p>\n After Cyclone Idai hit Mozambique in 2019, WFP\u2019s rapid disaster mapping operation deployed drones for the first time, allowing damage assessments in days rather than weeks.<\/p>\n \u201cIn just three days, locals were trained to use 10 different drone systems to capture more than 17,000 high-resolution images to inform relief operations,\u201d explained Annalisa Conte, Director of the Geneva office of the World Food Programme (WFP).<\/p><\/blockquote>\n A digital engine for emergency photo analysis (DEEP) also made its debut there, using high-resolution images from satellites and drones to identify damaged buildings with 85 per cent accuracy within hours, lowering field operation costs.<\/p>\n The system appeared again in Colombia and the Philippines, as well as after the August 2020 port explosion in Beirut, Lebanon.<\/p>\n Maximize impact, minimize risk<\/strong><\/p>\n Neural networks need a lot of training data to continually improve their interconnected AI systems \u2013 more than is available either in space or in disaster areas or conflict zones, said Wedler of DLR\u2019s Projekt AHEAD.<\/p>\n WFP tackles this data scarcity through a global network of call centres, which collect information on food supply and hunger daily. The resulting data feeds into predictive machine learning models, which generate real-time forecasts.<\/p>\n Using this global hunger monitoring system, the organization can predict the magnitude and severity of hunger across 90 countries at any given time, said Conte. The resulting real-time forecasts enable humanitarian teams to address all forms of hunger more effectively, she added.<\/p><\/blockquote>\n![\"\"](\"https:\/\/www.fie.undef.edu.ar\/ceptm\/wp-content\/uploads\/2022\/08\/AHEAD-project-globus.jpg\")
<\/p>\n