Antibiotic resistance in bacteria is a rapidly escalating global threat, with projections estimating up to 2 million deaths annually from drug-resistant infections by 2050. As a result, innovative, non-antibiotic based approaches to isolate resistant bacteria are urgently needed. Several studies have reported that changing the morphology of bacteria subsequently impacts its motility, and this relationship may offer new strategies to combat antibiotic-resistant strains. However, the relationship between bacterial morphology and motility in different porous mediums remains poorly understood. Certain antibiotics, such as ampicillin and cephalexin, are known to induce elongation in susceptible E. coli cells. These elongated cells exhibit different motility patterns from their shorter, resistant counterparts, including wider turning angles and more linear trajectories, resulting in larger root-mean-square displacements (RMSD). In this study, we investigated whether these differences in motility could be exploited in porous or geometrically complex microenvironments to spatially separate resistant and susceptible bacteria. Using genetic engineering, we cultured E. Coli with ampicillin and L-arabinose to selectively elongate susceptible cells, and injected them into microfluidic chips with varying spatial geometric constraints. Then, through brightfield microscopy and Trackmate software, we analyzed motility patterns and quantified bacterial trajectories. Our findings indicate that differences in RMSD can be leveraged to passively concentrate non-elongated, antibiotic-resistant bacteria near the injection site. This microfluidics-based approach can offer a simple, low-cost method to physical separation based on motility, with future applications in diagnostics, bacterial strain identification, and the development of targeted antibiotics.