Strathmore University is at the forefront of leveraging advanced technology to address critical environmental issues. Dr. Henry Muchiri, a researcher at Strathmore University’s School of Computing and Engineering Sciences (SCES), has been instrumental in a ground-breaking project that uses artificial intelligence (AI) to combat illegal logging in Kenyan forests. This initiative, recognized with the 2023 Innovation Grants Program award by Climate Change AI, involves a collaborative effort with researchers from Université de Neuchâtel, the Kenya Forestry Research Institute, and the University of Huddersfield.

The project, dubbed ‘Curbing Illegal Logging Patterns Using Sound-based Detection Techniques,’ employs IoT devices equipped with AI algorithms to detect the sounds of logging tools in forests where such activities are prohibited. The real-time alerts generated by these devices aid forest rangers in taking immediate action to prevent illegal logging, thereby contributing to climate change mitigation and promoting sustainable forest management.

In this interview, Dr. Muchiri delves into the technical workings of AI and IoT in detecting illegal logging, the significance of interdisciplinary collaborations, and the project’s impact on local communities and the environment. He also discusses the challenges of data scarcity and infrastructure limitations in AI deployment for climate action and shares his vision for the long-term benefits of this innovative approach to forest conservation.

How does AI and IoT technology work in detecting illegal logging activities?

IoT is used to develop devices equipped with powerful acoustic sensors that can read sound signals of cutting tools such as pangas and chainsaws. These signals can then be transmitted to an AI model which is able to classify different sound signals from animals to vehicle engines. A positive indication of illegal logging would be the detection of a cutting tool originating from a forested region where logging is not permitted. These can be used to prepare real-time alerts with the physical location of the site of logging to the ground forest rangers for further action. We also seek to expand this further with drone technology, to determine how it can support operations of the Kenya Forest Service in handling such cases. This will resolve issues such as inaccessibility and limited equipment and personnel.

How have collaborations with institutions like the Université de Neuchâtel, Kenya Forestry Research Institute, and the University of Huddersfield contributed to the project?

The partner agencies of this project have been greatly instrumental in supporting the primary research team at Strathmore. At Strathmore University, the project is handled under the SCES Makerspace Lab, an upcoming research lab on technology and engineering. In this project, we have engaged both faculty and students from undergraduate and master’s programs. Our team includes Trevor Kingangi (BSc EEE), Ali Sheikh (BSc ICS, Class of 2024), Monique Wamona (BSc EEE), Sandra Kahoro (MSc IT), Allan Vikiru (MSc IT, Class of 2024), Zainabu Muti (MSc IT, Class of 2024), Daniel Simiyu (MSc IT, Class of 2024), Jamlick Ntonjira, and Solomon Kamau (MSc IT, Class of 2023). I am serving as the project co-investigator, leading this dedicated team. We have been focusing on development of the IoT architecture as well as the machine learning model for acoustic classification.

To support in IoT development, the team partnered with the Centre for Efficiency and Performance Engineering at the University of Huddersfield, led by Professor Wenxian Yang as project co-investigator, and researchers Dr Zuolu Wang and Samuel Ayankoso. The Laboratory of Conservation Biology at Université de Neuchâtel led by Professor Clara Zemp as project principal investigator and Kenya Forestry Research Institute led by Dr Anthony Macharia as project co-investigator, have been essential in establishing a connection between illegal logging and climate change. Their research on the effects of biodiversity on the functionality of ecosystems has assisted in establishing the degree at which climate change has affected dependents of communities impacted by illegal logging. This team has featured Caleb Ngaba (Université de Neuchâtel), Magnus Onyiriagwu (Université de Neuchâtel) and Leley Nereoh (Kenya Forestry Research Institute).

How does your project address the challenges of data scarcity and infrastructure limitations in AI deployment for climate action?

A major goal with this work is to prepare and publish datasets of environmental sounds as well as those of cutting tools. This will enhance future research in development of similar machine learning models for sound classification in forested environments. We also anticipate scaling the IoT architecture under development to create potentially low-cost real-time monitoring systems that can be deployed in forests across Kenya and Africa. Lastly, we anticipate applying drone technology to capture spatial data on above ground carbon stocks in forests. This can be used to assess and establish trends in carbon emissions that are essential in monitoring the effectiveness of climate change mitigation strategies.

How do you envision this project impacting the local communities and the environment in Kenya?

The direct beneficiaries of this project are the Kenya Forest Service and forest-dependent communities. The technologies we are developing will complement their efforts in forest management and conservation. By deploying and monitoring the architecture for logging sound detection, the Service will get to determine and act upon instances of encroachment with minimal effort. The forest dependent communities represented by community forest associations, are set to realise major economic and sociocultural gains. Reduction in illegal logging enhances resources for household usage such as wood fuel and preserves aspects of their culture such as the use of medicinal plants.

So far, testing has been conducted at Nyangores Forest in Bomet County, as well as Oloolua and Ngong Hills Forests in Kajiado County. The Forest Services of both regions and the Community Forest Association at Nyangores have continually expressed their appreciation for the technology in forest monitoring and conservation. In both regions, both groups have been greatly involved in the development and implementation of this study.

What long-term benefits do you foresee for forest management and climate change mitigation as a result of this project?

In the long run, we foresee a grand expansion of this technology to forests across the country and beyond as illegal logging and consequent deforestation is a prevailing concern in the fight against climate change. Evident in Kenya, this research aligns well with various policies for local and national development, such as The Third Medium Term Plan of Kenya’s Vision 2030. This includes the Forest Conservation and Management Act of 2016 and the National Climate Change Action Plan 2013 – 2017. These policies highlight the development of technologies for restoring and monitoring forests at both national and county levels.

Beyond supporting forest conservators in their operations, forest-dependent communities are set to benefit from the dissemination of this study. This can be through re-evaluation of consumption of forest-derived products such as charcoal and seek to adopt climate-friendlier sources for household usage. Additionally, communities can develop programs such as agroforestry, enabling youth and women’s groups to achieve economic empowerment, as reported in other existing programs across Kenya. Besides these initiatives, researchers in this study aim to contribute to and expand the research into areas such as biodiversity conservation, sustainability, bio-acoustic machine learning, edge computing, and energy-efficient IoT.

What has been the most rewarding aspect of working on this project so far?

A major inspiration from this work has been the spirit of collaboration and camaraderie among members of the research team and external stakeholders. People of different nationalities, cultural and educational backgrounds, genders, and ages have come together with the common goal of conserving and protecting forests. Young people, who constitute over 80% of the research team, have particularly benefitted from this study. By participating directly in cutting-edge research on the climate crisis, a topic of great importance to them, they are contributing to a cause crucial to the future of our planet.

Article written by Stephen Wakhu