Tag Archives: Local

Humanitarian Robotics, Murphy’s Law and What To Do About It

Like any other technology used in humanitarian settings, robotics solutions can break down when you need them the most. A few months ago, for example, my team and I at WeRobotics were in the middle of the Peruvian Amazon Rainforest with a relatively expensive cargo drone that could hardly fly without become dangerously unstable. Murphy’s law is alive and well in the Amazon as it is in other places we work in like Tanzania, Nepal, Haiti and Maldives. So what to do?

Introducing emerging technologies in aid and development projects in the global South comes with a range of challenges and responsibilities. What’s the point of transferring robotics solutions to local partners if these platforms break and can’t be repaired locally? In one country we work in, for example, a major international organization has purchased about a dozen flying robots, and every few months at least one of these UAVs has to be shipped back to Europe for repairs. Not only does this really add up in terms of shipping costs, but it also creates significant project delays when half your fleet is out of the country for months on end. 

In Nepal last year, our Flying Labs team were out of propellors which meant we had to ship some new ones in from Europe. This is expensive and it didn’t work: the propellors were returned to us 2 months later because the shipping service had not found the address of our local Flying Labs Coordinator. (Yes, we’re exploring 3D printer solutions, but these break as well). In Tanzania, the UAV pictured above has seen a frustrating number of technical and software failures, which has prevented our Flying Labs from actually completing important projects. That particular UAV has had to be shipped back to Europe twice for repairs, costing both time and money.

So what to do? Going with cheaper, “DIY” UAVs doesn’t necessarily solve the issue. These don’t tend to be as robust or easy to use even if they are more expendable than costly models. That said, the most expensive UAV in our Flying Labs fleet has been the most problematic in terms of repeated technical failures. Sure, we could buy more reliable (costly) UAVs and have backups just in case but this does require more funding, and these UAVs will inevitably require repairs at some point too. So this “solution” doesn’t actually address the underlying issue: the dependency we create when introducing these new robotics solutions.

Obviously we need to train our Flying Labs to repair and service these UAVs locally. We’ve started doing this, and while our Labs won’t become maintenance maestros overnight, I’m personally really excited that we’re moving forward on this. Instead of shipping UAVs back to Europe for repairs, we’ll eventually be able to repair most technical problems onsite at our Tanzania Flying Labs, for example. Besides the obvious advantages (cost-savings and time-savings), this service will generate an important source of income for our local Flying Labs staff. And given that the mandate of our Labs is to create local jobs and incubate local businesses that offer robotics as service, one such business could well specialize in repairs and maintenance. 

So when international organizations and companies in the country or region in question need their UAVs fixed, they could pay our Labs to carry out repairs instead of shipping then back to manufacturers in Europe or the US. There is a small catch, however. By repairing the UAVs ourselves, we run the risk of voiding the warranty on the UAV. So we’re starting with small, common repairs that don’t pose this problem. But in the long run, we want to have leading UAV manufacturers certify our Flying Labs as official partners for repairs. This too won’t happen overnight. First we first need to prove ourselves with basic repairs and clearly demonstrate the savings in cost and time that UAV operators gain from having their UAVs fixed at one of our local labs.

We’re heading back to Tanzania in a few weeks to provide additional training on how to repair these technologies locally. If you’d like to help us train our Flying Labs on UAV/drone repairs and maintenance, please do get in touch. Thanks!

Think Global, Fly Local: The Future of Aerial Robotics for Disaster Response

First responders during disasters are not the United Nations or the Red Cross. The real first responders, by definition, are the local communities; always have been, always will be. So the question is: can robotics empower local communities to respond and recover both faster and better? I believe the answer is Yes.

But lets look at the alternative. As we’ve seen from recent disasters, the majority of teams that deploy with aerial robotics (UAVs) do so from the US, Europe and Australia. The mobilization costs involved in flying a professional team across the world—not to mention their robotics equipment—is not insignificant. And this doesn’t even include the hotel costs for a multi-person team over the course of a mission. When you factor in these costs on top of the consulting fees owed to professional international robotics teams, then of course the use of aerial robotics versus space robotics (satellites) becomes harder to justify.

There is also an important time factor. The time it takes for international teams to obtain the necessary export/import permits and customs clearance can be highly unpredictable. More than one international UAV team that (self) deployed to Nepal after the tragic 2015 Earthquake had their robotics platforms held up in customs for days. And of course there’s the question of getting regulatory approval for robotics flights. Lastly, international teams (especially companies and start-up’s) may have little to no prior experience working in the country they’re deploying to; they may not know the culture or speak the language. This too creates friction and can slow down a humanitarian robotics mission.

WeR PM Blog pic 2

What if you had fully trained teams on the ground already? Not an international team, but a local expert robotics team that obviously speaks the local language, understands local customs and already has a relationship with the country’s Civil Aviation Authority. A local team does not need to waste time with export/import permits or customs clearance; doesn’t need expensive international flights or weeks’ worth of hotel accommodations. They’re on site, and ready to deploy at a moment’s notice. Not only would this response be faster, it would be orders of magnitudes cheaper and more sustainable to carry through to the recovery and reconstruction phase.

In sum, we need to co-create local Flying Labs with local partners including universities, NGOs, companies and government partners. Not only would these Labs be far more agile and rapid vis-a-vis disaster response efforts, they would also be far more sustainable and their impact more scalable than deploying international robotics teams. This is one of the main reasons why my team and I at WeRobotics are looking to co-create and connect a number of Flying Labs in disaster prone countries across Asia, Africa and Latin America. With these Flying Labs in place, the cost of rapidly acquiring high quality aerial imagery will fall significantly. Think Global, Fly Local.

Counter-Mapping the State with UAVs

Want a piece of Indonesia? The country’s government is busy implementing an “accelerated development program” in which “different provinces are assigned different development foci,” like “food and energy for Papua, palm oil processing for North Sumatra, mining for Central Kalimantan etc.” Critics describe this program as “a national, state-coordinated program of land grabs.” An important component of “this development plan is the commoditization of space by spatial planning,” which is “supposed to be open, transparent and participatory.” The reality is very different. “Maps are made by consultants and government offices favoring the interests of capital and local elites.” As a result, “concessions are given mostly without the consent (and often without the knowledge) of local communities.” These quotes are taken from a brilliant new study (PDF) written by Irendra Radjawali and Oliver Pye. The study describes the use of Unmanned Aerial Vehicles (UAVs) to “generate high-quality community controlled maps to challenge spatial planning from above,” which is “revolutionizing the counter-mapping movement in Indonesia.”

Screen Shot 2015-07-31 at 2.46.00 PM

“Challenging state power over maps and its categorization of land uses by counter-mapping indigenous and local claims to territory has developed into an important movement in Indonesia.” As the authors of the new study rightly note, “Mapping needs to be understood as a political process rather than a merely technical tool. Mapping is not only an act of how to produce maps, it is important to always ask who produces the maps, how people can access the maps and how the maps can be used for emancipatory purposes.” Counter-mapping is thus a political process as well. And this counter-mapping movement is now experimenting with “grassroots UAVs” (or community drones) to bolster their political actions.

Activists in Indonesia initially used their UAV to capture “high quality and high-resolution spatial data in areas where access was restricted by company security and police.” Where exactly did they get their UAV from? They built one from scratch: “Irendra Radjawali built the first drone without any former training, by using the Internet and the online forum. He also sourced much of the material second hand via ebay.” The advantage of this DIY approach is the relatively low costs involved. This UAV, coupled with a mapping camera, came to just over USD 500.

Screen Shot 2015-07-31 at 3.35.58 PM

Irendra and team subsequently few their UAVs over oil palm plantations where a company had taken lands from local communities who had no idea that their lands had been parceled off to said company. The team managed to fly their UAVs “at several places, capturing several community’s areas which have been grabbed by the company, including the customary area.” It is worth emphasizing that “community members very rarely have access to the spatial plan documents, and so could hardly ever actively participate in the spatial planning process. The opportunity to produce high-quality and precise maps is seen by community members as the chance to claim and to re-claim their lands.”

The team also flew over an area that was directly “affected by the expansion of large scale open mining for bauxite.” The water from the river became unsafe to drink; fishing grounds vanished; the nearby lake dried up. Local communities repeatedly protested the irreversible destruction of their ecosystem but this hasn’t stopped mining companies from expanding their activities. Irendra and team were able to take aerial photographs of the affected areas. One of the “high-quality and precise maps” that they were able to generate with these photos has since “been used as an evidence to disclose illegal mining company exploiting bauxites operating outside of their concession area.” These aerial counter-maps are thus “being used to provide evidence against the mining company,” and they also support local community’s efforts to protect their existing lands and forest.”

Screen Shot 2015-07-31 at 3.45.19 PM

Irendra and his colleagues took a direct, community-driven approach to these counter-mapping projects: “Community members are involved in establishing the community drone and in deciding who will be responsible to perform drone mapping activities. […] Village meetings also discussed the plans and strategies to perform mapping activities at various different villages with different challenges and contexts. One part of village meetings was training on mapping and drones where participants were informed about participatory counter-mapping techniques as well as the use and the operation of drones to support rapid participatory counter-mapping for high-quality spatial data. A meeting in Subah village agreed to fund the mapping themselves by a monthly contribution of [50 USD] from each [sub-village].”

In sum, co-authors Irendra and Oliver write that UAVs are “very empowering.” “The sense of power and achievement when community members themselves fly the drone is substantial. The empowerment impact that comes with the knowledge that these images are of greater quality than the concession maps and that they have been acknowledged by the Constitutional Court is even greater.”

It is worth noting that the land-use planning maps controlled by the government and companies were made on “the basis of satellite imagery,” which means that “small hamlets [are] not visible. In the process of map-making by the State, the hamlets literally disappeared, losing any rights to their land in the process. With high-resolution drone maps, however, residential areas, farming, fruit tree forests and other long-term uses of the land are rendered visible. Furthermore, local communities require high-quality maps to re-claim those residential areas which now are ‘officially’ part of company’s concessions. These maps are used to support their arguments to halt new concessions for mining and for oil palm.”

Screen Shot 2015-07-31 at 4.39.47 PM

Not surprisingly, perhaps, “the counter-mapping process also uncovered simmering territorial conflicts.” In one of these conflicts, “it emerged that the unsettled village border is one the problems.” Irendra and fellow co-author Oliver write that “One of the aims of community drones is to map the area of several villages […] and to confirm village borders.”

The team’s use of UAVs for counter-mapping resulted in a number of political victories that went beyond the local level. In one case, for example, a counter-map was “used as legal evidence at the Constitutional Court trial on the 1st September 2014, providing the chance for drone counter maps to be recognized by the Indonesian legal system in the future.” In another case, counter maps were combined with other evidence to “challenge the provincial government to accept what the civil society organizations demand. Some of their demands were achieved and accepted, including: (1) Recognition of community-managed lands, (2) Recognition of customary community rights, and (3) active community engagement in the spatial planning process. These demands had not been addressed before.” In yet a third case, “Maps made by drones were used to support […] arguments that often mining activities are causing detrimental social and ecological effects. The Constitutional Court ruled against the mining corporations [as a result], upholding the obligation of mining companies to install smelters and to process raw minerals and coal before exporting them.”

Screen Shot 2015-07-31 at 4.37.53 PM

These projects have generated a growing interest in UAVs, which is why the local Swandiri Institute recently established a “drones school” where “civil society organizations and community activists who are interested in learning and using drones for mapping and for advocacy work could join and participate.” A second drones school was also launched by other partners to “focus on using drones at village levels to map village areas and to confirm village borders.”

The authors conclude that “the appropriation of drone technology by community activists has the potential to improve the situation with regard to inclusion, transparency, and empowerment. […] Nowadays, younger members of local communities are computer literate. After a mapping flight, images and videos can be directly downloaded on to a laptop, giving instant transparency to village meetings during the mapping project. The resolution is so high that individual houses, trees, etc. can be clearly identified, also increasing transparency and the potential to include just about everybody in territorial discussions.”

But of course, to state the obvious: UAVs are not a silver bullet or “magic wand that can conjure away hierarchies and power structures at the local level or in wider society.” Irendra and team were “unable to use drones in those areas where local elites were in cahoots with plantation and mining companies and controlled traditional institutions such as customary councils and where opposition was marginalized.” In other areas, “hierarchical gender relations […], power dynamics, and territorial disputes between different villages were replicated in the mapping process.” At the same time, the UAV revolution does have “the potential—together with campaigning and political pressure—to force through the recognition of community counter-maps in the spatial planning process […].” To this end, “if embedded within political action, drone technology can revolutionize counter-mapping and become an effective weapon in the struggle against land grabs.” And in this context, “community drones for counter-mapping could well become a technology of the masses, by the masses, and for the masses.”

Automatically Identifying Eyewitness Reporters on Twitter During Disasters

My colleague Kate Starbird recently shared a very neat study entitled “Learning from the Crowd: Collaborative Filtering Techniques for Identifying On-the-Ground Twitterers during Mass Disruptions” (PDF). As she and her co-authors rightly argue, “most Twitter activity during mass disruption events is generated by the remote crowd.” So can we use advanced computing to rapidly identify Twitter users who are reporting from ground zero? The answer is yes.

twitter-disaster-test

An important indicator of whether or not a Twitter user is reporting from the scene of a crisis is the number of times they are retweeted. During the Egyptian revolution in early 2011, “nearly 30% of highly retweeted Twitter users were physically present at those protest events.” Kate et al. drew on this insight to study tweets posted during the Occupy Wall Street (OWS) protests in September 2011. The authors manually analyzed a sample of more than 2,300 Twitter users to determine which were tweeting from the protests. They found that 4.5% of Twitter users in their sample were actually onsite. Using this dataset as training data, Kate et al. were able to develop a classifier that can automatically identify Twitter users reporting from the protests with an accuracy of just shy of 70%. I expect that more training data could very well help increase this accuracy score. 

In any event, “the information resulting from this or any filtering technique must be further combined with human judgment to assess its accuracy.” As the authors rightly note, “this ‘limitation’ fits well within an information space that is witnessing the rise of digital volunteer communities who monitor multiple data sources, including social media, looking to identify and amplify new information coming from the ground.” To be sure, “For volunteers like these, the use of techniques that increase the signal to noise ratio in the data has the potential to drastically reduce the amount of work they must do. The model that we have outlined does not result in perfect classification, but it does increase this signal-to-noise ratio substantially—tripling it in fact.”

I really hope that someone will leverage Kate’s important work to develop a standalone platform that automatically generates a list of Twitter users who are reporting from disaster-affected areas. This would be a very worthwhile contribution to the ecosystem of next-generation humanitarian technologies. In the meantime, perhaps QCRI’s Artificial Intelligence for Disaster Response (AIDR) platform will help digital humanitarians automatically identify tweets posted by eyewitnesses. I’m optimistic since we were able to create a machine learning classifier with an accuracy of 80%-90% for eyewitness tweets. More on this in our recent study

MOchin - talked to family

One question that remains is how to automatically identify tweets like the one above? This person is not an eyewitness but was likely on the phone with her family who are closer to the action. How do we develop a classifier to catch these “second-hand” eyewitness reports?

bio