A photoplethysmogram (PPG) sensor has been broadly used for smart watches and bands because it is easy to measure and contains many health information, such as heart rate (HR) and respiratory rate (RR). Because the PPG sensor blinks an LED at sampling instants, the rate for sampling and LED flashing should be reduced to extend the battery life. To reduce the rate, we employed two different compressive covariance sensing (CCS) techniques and applied them to HR and RR estimation. The CCS cannot recover a signal itself but reconstructs its covariance. We designed a signal processing technique to extract HR and RR from the reconstructed covariance. The estimation performance was evaluated by using the open-source data and the experimental data, and the power consumption of a wrist-type PPG sensor with respect to the compression ratio was evaluated. The proposed method acquired the covariance of the PPG with an average sampling rate of 1.79 Hz below the Nyquist rate for HR (10 Hz), and it significantly reduced the energy consumption for the PPG sampling. Moreover, its estimation accuracy was sufficient to be used for a wearable healthcare system. As a result, the proposed method showed that HR and RR can be estimated with an ultra-low-power PPG sensor. INDEX TERMS Compressed sensing, health information and management, energy efficiency, photoplethysmogram. The associate editor coordinating the review of this manuscript and approving it for publication was Sabah Mohammed.