Energy harvesting offers the promise of mobile sensor systems capable of quasi-perpetual operation, but the discontinuous and dynamic characteristics of harvesting in real-world scenarios - necessary for the design and operation of self-powered systems - are not yet well understood. The paper presents a hardware platform for providing a comprehensive real-world evaluation of two energy harvesting modalities common to body sensor networks: indoor light and thermoelectric. Day-long multi-modal energy harvesting profiles were generated, which were then used to develop a mathematical model to predict real time energy harvesting values from the sampled environmental and human behavioral parameters. Experimental results demonstrate that the model is effective in calculating and predicting harvested energy in real time, and a multi-source scheme for continuous operation of self-powered sensors is demonstrated.