In recent years, there has been growing interest in understanding, measuring, and modeling the in-vehicle workload of the driver. Drivers sometimes engage in potentially distracting activity while driving because they believe they can do so without compromising the driving operation. An accurate and real-time driver workload measure would be useful in the design of vehicle safety systems such as crash avoidance, and the system which controls the transition from automated to manual driving mode, and vice versa. There are several important types of in-vehicle driver workloads: (i) Driver-vehicle control action, (ii) Driver-instrument-panel interaction, and (iii) Inter-vehicle prediction in traffic environment. In this paper, we describe an in-vehicle data acquisition system, TIP_WL, and set of protocols for capturing these three types of driver workload. The system consists of three major modules: a physiological module, a vehicle module, and an event recording mobile application. The physiological module includes physical sensors which the driver wears. The module collects real-time physiological data about the driver, including ECG signals, heart rate, respiration rate, and skin conductance. The vehicle module includes an OBD vehicle data acquisition system, a GPS recording device, and two video cameras. The mobile app module is developed to record traffic events (turns, lane changes, traffic signs), as well as driver workload level. Note that a human expert operates the mobile recording app while the driver drives the vehicle. Using a 3-level scale (low, medium, high), the following information is recorded for each event: the expert’s estimate of the driver’s workload, the driver's estimate of their own workload, and the density of traffic on the road. To assess driver instrument panel interaction, a script is used whereby at random times during the trip, the driver receives a phone call with instructions on what actions to perform in the vehicle, such as: turn the radio on or off or to a different station, adjust the air conditioning or heat, etc. In this paper, we explain how the mobile app was designed and show examples of the types of data that have been collected. We will also present preliminary results on the performance of the driver workload estimation algorithms that we developed.