The future of insect farming is happening now
According to Microsoft’s “IoT Signals” report released in 2019, approximately 30% of IoT deployment initiatives fail in the proof-of-concept phase, while 38% of IoT adopters said that the technical challenges and complexity of IoT hinder their ability to move forward. Our quest to provide a comprehensive proof of concept on the use of IoT sensor technology for optimal insect-based animal protein production is on a path to defying these statistics — perhaps because our experiments are iterative and based on continuous learning. Right from the start of the project, we knew that IoT would be a game-changer with respect to unlocking the potential of insect farming in Africa. But we also knew that as we incorporate and test more IoT sensors in monitoring the key insect growth parameters, the system would become more complex and present unique challenges.
If you can find a path with no obstacles, it probably doesn’t lead anywhere
To realize its full potential, the IoT system utilizes many different resources and assets. These range from the sensors, connectivity and data analytics to user interface. The integration of these 4 components make IoT so powerful but at the same time so weak if one of the components is malfunctioning or working sub-optimally. To ensure that we generate the biggest bang for the buck and put the project on track to achieving the intended outcomes, we laid a bigger emphasis on turning Sprint 1 challenges into key success points that subsequent sprints would leverage as a springboard to achieving the desired results:
(a) Connectivity
A seamless flow of information between interconnected devices and the cloud is the entire basis of a successful IoT deployment. In Sprint 1, internet reliability presented a major bottleneck to the extent that there were some data losses mainly as a result of intermittent loss in internet connectivity. To deal with this situation, our technology partner, Wisense, spent a significant amount of time in Sprint 2 redesigning the sensors to ensure fail-safe data storage when internet connection is lost. This marks a big step especially in ensuring that the sensors are adapted to the needs of small-scale Black Soldier Fly farmers who are located in remote locations where internet connectivity is not strong enough and/or not reliable.
To maintain a reliable internet connection, and to maximize on the value of the IoT sensor technology, Sanergy is currently upgrading its internet infrastructure to a multiple access point network which delivers a strong and fast internet throughout. The multiple access points will blanket our coverage area with a mesh of internet connectivity so every area has a strong signal without any dead zones.
(b) Gateway/sensor distance range
Proper gateway placements have a great impact on IoT sensor technology performance. This was a key learning in Sprint 1 where the sensor nodes had been placed far away from the gateway resulting in poor connection. Gateways act as a bridge between the sensors and the cloud. In addition, the gateways can provide extra storage and processing services making the sensor nodes more power efficient and fast. In Sprint 2, additional gateways were introduced to reduce the communication distance between the sensor nodes and the gateways. Though monitoring is still ongoing, initial results show that having sensor nodes closer to the gateway effectively eliminates the challenge of poor connection. But these initial results also show that finding the optimal distance between sensor nodes and gateway will result in the most reliable and stable connection. Effective deployment of the gateways for stability of the entire sensor network will therefore form a key experiment in Sprint 3.
The development Journey
As it is important that challenges affecting IoT sensor technology reliability are solved as they occur, and before volume roll-out that includes several sensors, learning from Sprint 1 were critical in Sprint 2 with respect to advancing the sensor design and developing the desired mounting features which could easily be adapted to the requirements of small-scale BSF farmers. However, this redesign process faced several challenges as it required some research time and sourcing of highly amenable yet durable sensor components that were not locally available. As a result, a significant amount of Sprint 2 time went into the redesign and sensor development process leaving us with only a couple of days to deploy and monitor the sensors in the field. Initial results show that the time spent in the redesign process was worth it and will be beneficial in the long run when the sensors are deployed in small-scale BSF farms where it can be time and resource intensive to troubleshoot challenges experienced in Sprint 1. It is important to note that our experiments are iterative and that in every subsequent sprint, and based on the learning from the previous sprint, a component of the sensor or the entire network may need to be redesigned. In short, the redesign process could be repeated several times as dictated by new learning before the desired functionality is achieved. IoT sensor technology has already been proven and the bulk of our work through this project lies in confirming what works for both large and small-scale BSF farmers, suitably adapting the technology in a way that is inclusive and finding solutions to challenges that may accompany their uptake.
Your life could be an experiment. But not the only experiment
Experiments build an understanding about what works by destroying misunderstanding and errors. By revolutionizing the way food is produced, IoT sensor technology will transform the lives of small-holder farmers who produce as much as 75% of all food grown in sub-saharan Africa yet make up the majority of the world’s population living below the poverty line. But to realize this transformation, we have to be led by facts. Rigorous evidence on the potential of IoT sensor technology to advance processes and optimize food production regardless of the size of farms.
A total of 4 experiments were planned in Sprint 2 but only 3 happened. An experiment to monitor Ammonia levels in wet larvae growhouse did not happen because the design process of the Ammonia sensors went beyond the Sprint 2 timeline.
In one experiment, three light intensity sensors were deployed at Sanergy’s adult BSF rearing facility in Kilifi, Kenya and monitoring is currently ongoing. These sensors measure ambient illuminance by converting light energy into electronic signals. A study done by Wageningen University in 2019 suggests that mating success of reared Black Soldier Fly can be dramatically increased by exposing the adults to light that is particularly rich in wavelengths near 440 and/or 540 nm and has an irradiance that is an appreciable fraction of the intensity of full sunlight. Our preliminary results appear to confirm this suggestion. Conclusive observations will be made when monitoring of these sensors is completed.
Accurate and quantifiable measurement of light intensity, and ability to get this data in real-time will allow BSF farmers to respond rapidly to changing light intensity and optimize the growth conditions during the fast life-cycle of the target insects. Conclusive results will open the vast untapped potential for smallholder farmers as the ability to monitor and maintain optimal breeding conditions will increase egg production which directly translates to increased insect-protein production.
In another experiment, the software development work aimed at integrating the sensor dashboard with an Enterprise Resource Planning system was very successful and gave us some insights on how creating a unified, single view of all data would result in faster and better decision making. The software showed to be fit for purpose as it was able to aggregate and combine data from 4 different sensors and test them as a group. This integration work will open a new door of excellence in insect farming by creating multiple user access and making data available in real-time at different locations for timely decision making.
For the final experiment we conducted, Sanergy and the International Centre of Insect Physiology and Ecology (Icipe) came out for field visits to small-scale BSF farms to develop an understanding around how the sensors could be matched to the requirements of the small-scale farmers based on their existing farm infrastructure. In each farm, the team asked the farmers questions on their current processes, performance and future plans. Tour stops in each farm included waste receiving and pre-processing, egg production, larvae growing and product harvesting units. These visits confirmed that currently, small-scale BSF farming involves a lot of manual monitoring of the growth conditions through paper-based data collection and difficult data analysis during the first life cycle of the Black Soldier Flies. The small-scale farmers expressed strong interest in learning more about IoT sensor technology, data that could be beneficial to them and getting involved in the project.
The small-scale farms visited use stacked bins for BSF young larvae growing and adult cages are mounted on shelves. The project had visualized this kind of set-up in small-scale farms and the design work already done will be easily adapted in Sprint 3 where the first batch of sensors will be deployed in these small-scale farms.
Long story short
According to Action Against Hunger, 24.1% of the population of Sub-Saharan Africa is undernourished, the highest prevalence of all regions in the world. At the same time, 413 million people in the same region are living on less than $1.90 per day. These statistics are depressing. But there is an enormous opportunity. Following the completion of the first 2 Sprints, our belief that IoT sensor technology can cure the food security challenges in Africa and bring major economic and environmental benefits through optimized insect production has crystallized. We already know what will work and what won’t work for the small-scale farmers; the next phase of our experiments will have a strong bias for working with the small-scale BSF farmers to reliably deploy and monitor the sensors in their farms and demonstrate that the technology can be easily adopted by small-scale farmers as well to optimize BSF production. Although we do not think that the adoption of IoT technology among small-scale BSF farmers will come with the kind of disruptive revolution we saw with mobile phones in Africa, the dissemination of results, especially those derived from small-scale farms, will make this technology more practical and sensible for all small-scale insect farmers across Africa to take up.
