The presence of noise in an electrocardiogram (ECG) signal is detrimental to the accuracy of R-peak detection for the purpose of heart rate variability (HRV) computation. Electrode motion noise in particular is adverse to the continued development of many beneficial technologies involving noise robust ECG fiducial marker detection algorithms, specifically for mobile acquisition devices. This paper proposes a novel technique for R-peak detection utilising dynamic time warping (DTW) with improved tolerance to electrode motion noise compared with that of established techniques. The MIT/BIH Arrhythmia Public Database was used in the experiments, with several added SNR levels of noise ranging from -10dB to 20dB. Metric evaluation indicates the DTW detection algorithm exhibiting favourable performance in comparison to established techniques for SNR levels below 10dB with a sensitivity of 95.93% and positive predictability of 91.73%. Performance degradation due to increased noise power is similarly favourable to the DTW detection technique in comparison to benchmarks, indicating the noise robust capability. The technique also has merit as a viable competitor for QRS detection in non-noisy ECG signals, with a sensitivity of 97.21% and positive predictability of 94.46% on the clean dataset.