Accurate indoor localization requires non line-of-sight ranging error identification and mitigation. Current research has focused on methods that improve accuracy but are too complex to be executed in resource-constrained hardware in real-time. In this paper, we propose and evaluate a non line-of-sight mitigation technique using impulse radio ultra-wideband (IR-UWB) experimental ranging data. We show that using this technique, ranging non-line-of-sight (NLOS) errors can be significantly reduced in the location estimate. The method is less complex than other methods used in related work and can potentially be implemented on microcontrollers.