How machine learning could make drones more practical
Automation has been essential to the long-term development of every industry from manufacturing to accounting. Assembly lines, software and now autonomous machines are just a few forms that automation has taken over the years. Although most drones currently require some degree of human control, they could become more independent in the future thanks to machine learning capabilities. Accordingly, they could become more widely used in mines and in public safety capacities such as firefighting.
Drones and machine learning: What are the possibilities?
Machine learning has become one of the biggest buzzwords in tech in recent years. The concept essentially encompasses the use of computers that can learn concepts without having been explicitly programmed to perform them. According to Research and Markets, the global cognitive computing market, inclusive of machine learning, was worth $2.5 billion in 2014 but was expected to balloon to $12.5 billion by 2019, for a compound annual growth rate of 38 percent.
But what exactly is machine learning? Chances are that you may have been affected by a machine learning application without even knowing it. Some common examples include:
- Web search engines: These applications return recommendations as well as what they deem “relevant” results, based on the input they get over time.
- 3-D modeling: Maps and charts are often built with modeling techniques that automatically pull information from photographs.
- Voice assistants: Popular interfaces such as Apple Siri, Google Assistant and Amazon Alexa can supply natural language responses to user queries and learn from habits.
Machine learning covers a broad range of use cases, many of which may be useful to drones, especially as manufacturers begin outfitting their models with more sophisticated CPUs, batteries and cameras. Chip maker Qualcomm, one of the industry leaders in the creation of systems-on-a-chip for mobile phones, showed off its Snapdragon Flight Drone Platform at the 2017 Consumer Electronic Show in Las Vegas.
Drones could become more autonomous with the help of machine learning.
In the real world, this solution would enable drones to respond to unexpected conditions, such as walls or inclement weather, and chart new paths. Such autonomous operation would be particularly important for drones placed in settings such as mines, in which minimizing the number of human personnel is often a safety priority.
Drones: The new canaries in the coal mine
The saying “canary in the coal mine” has its roots in the practice of lowering a canary into a coal mine to test the air for concentrations of methane and carbon dioxide, two gases to which this bird species is particularly sensitive. Drones could become mechanical canaries in the years ahead by providing the photography needed to map mine sites, if not test their air supplies.
Automated gathering of site data by drones could provide a detailed picture of each mine, without putting humans at risk of injury or the long-term exposure that heightens the risks of diseases such as black lung. As machine learning and drone hardware both advance, expect drones to pop up in many new settings, from mines to golf courses, to assist with mapping and various other uses cases.
