Path Planning in Informal Settlements

The motivation for this work relates to the growing issue of rapid urbanization in the expanding communities of South Africa. To address the immediate demand for
shelter, numerous communities established housing structures independent of the government. These communities lack sufficient infrastructure for municipal services, compounding the damages done by emergencies such as fires, floods,
assaults, and illness. We can model the issue of traversing these communities to deliver crucial resources as a shortest-path problem: with a singular agent attempting to navigate a cluttered environment to reach a goal state subject to a global constraint.

As a 2022 Robotics Institute Summer Scholar in the Search-Based Planning Lab, I drastically honed my proficiency in C++ and cultivated new algorithmic thinking patterns to improve the living and safety conditions within the informal settlements of South Africa. I tested the performance of three variations of the A* path-planning algorithm to discern the optimal performance in congested urban environments. These algorithms were realized in C++ using hash-maps to ensure computational efficiency. The environments were modeled with 2D maps of urban cities from the Moving AI benchmark where I applied a first-order Gaussian filter to randomly simulate the agent’s discrete traversal cost. In this summer internship at Carnegie Mellon University, I earned invaluable experience merging my technical discipline with my social consciousness, informed from my study abroad experience in South Africa earlier in the year. This experience begins to signify the direction of my professional career, one intertwined in novel robotic applications to serve minoritized groups.