The precision race car: IoT and the story of Sam Schmidt

By | 6 minute read | October 11, 2016

Race car

A few years ago I was introduced to track racing and have since had the opportunity to drive at Laguna Seca where I navigated the infamous Corkscrew—successfully. So when I heard Sam Schmidt’s story, I was impressed, to say the least. Sam started racing professionally at the age of 31—which is relatively old compared to professional racing norms—and was named rookie of the year in 1995. Considered a rising star, Sam raced three consecutive times in the Indianapolis 500 and netted his first victory from the pole at the Las Vegas Motor Speedway in 1999.

In January 2000, Sam had a terrible crash during a practice lap at the Walt Disney World Speedway in Orlando, FL and was tragically paralyzed from the neck down. Sam’s racing days were over, or so it seemed.

In 2013, Sam met with a group of engineers at Arrow Electronics Inc. and decided to set course with an ambitious goal: enable Sam to drive again.  By 2014, Sam was speeding down the racetrack at more than 100 mph—driving all on his own. Read on to find out how.


Figure 1: High Level SAM Car Architecture

Figure 1: High Level SAM Car Architecture

In 2014, the engineering team at Arrow created the first version of a semi-autonomous motorcar (SAM) car. The current version of the SAM Car debuted in early 2016 and is a modified 2016 Corvette ZO6. The heart of the car’s systems is the SAM Computer, which provides the gateway between a number of cameras and sensors, the actuator control system that controls the vehicle, and a GPS unit for precise, real-time monitoring of the car’s physical location.

The SAM Computer gateway is from Advantech. It has been ruggedized for automotive applications and runs on an Intel Atom-based processor and the QNX operating system. It also runs the Arrow Connect gateway software.

The SAM Car’s sensors pull data associated with both the car and the car’s driver. The car data includes information on the car’s performance, handling and speed, along with the in-vehicle environment such as brightness, temperature and humidity. Sensors also gather data about the driver such as heart rate and body temperature.

Figure 2: Sensors embedded into the driver’s helmet

Figure 2: Sensors embedded into the driver’s helmet

To handle steering, there are eight reflective, infrared sensors fitted into the driver’s helmet, while four infrared cameras face the driver. To steer, the driver simply has to look in the direction he or she wants to go. For you track drivers, that would be the apex of the turn (usually marked with a bright orange cone for us rookies).

All input from the cameras and sensors integrate into a single camera PC , which connects to the SAM Computer to track the driver’s head movements in real time. The camera PC uses an Intel i7 processor running Microsoft Windows 7. The system also deploys Motive software that tracks the rigid body markers on the driver’s helmet or sunglasses and outputs the angle of the driver’s head from the camera PC to the SAM Computer.

Figure 3: Acceleration and braking via the driver’s mouthpiece

Figure 3: Acceleration and braking via the driver’s mouthpiece

To handle acceleration the driver puffs breaths of air into a mouthpiece equipped with an NXP pressure sensor, specifically selected to be sensitive enough to respond to the driver’s breath input: the stronger the breath intensity, the stronger the acceleration rate. The car then responds directly via a rotary actuator attached to the gas pedal. The gas pedal is depressed based on the amount of air pressure the driver creates, giving him full control over acceleration, from a smooth gradual increase to the quick acceleration demanded in racing conditions.

Braking is accomplished through the same mouth-pressure sensor. The driver sips on the mouthpiece creating negative pressure that the system translates into braking. The driver can coast at a chosen speed by not continuing to sip or puff air after the desired speed is achieved.

The actual control of the car is handled by a Paravan  drive system, which includes microprocessors that transmit signals in nanoseconds to one servomotor for the brake and accelerator and a rotary actuator on the steering wheel. Sensor and camera data processed by the SAM Computer are sent as messages over a controller area network to the Paravan system. The Paravan drive system sends updates 100 times per second (10 milliseconds) to the actuators enabling real-time control of the car.


Through a GPS unit located onboard the vehicle, the SAM Computer connects outside of the car over a 4G LTE modem to the Arrow Connect Cloud.

One form of data transmitted is the vehicle’s telemetry data, using TLS 1.2 for transport, AES-256 for data encryption and SHA-256 for data hashing. This data is used to keep the car within 1.5 meters from the inside edge of the racetrack wall. In essence, this GPS system establishes “visual curbs” or boundaries, as the car is programmed to act according to its GPS location. If the car gets too close to the wall, it warns the driver to correct course. If the car continues to drift toward the wall, the system gently auto-corrects the car to keep it safely on the track.

In addition, the car’s sensor and video data are also transmitted. Sensor data is streamed in real-time—or stored and forwarded when in a disconnected state—and is transmitted using REST and MQTT protocols. Also, when high-speed Internet connectivity is available, the SAM Computer stores and forwards the video.


The Arrow Connect Cloud runs on the Microsoft Azure compute & storage cloud service and uses MongoDB as its underlying database. To date, the SAM Car has generated and stored more than 5 million records that include 1GB of sensor data and 3GB of videos. This data was generated from about one dozen races/events.


Figure 4: SAM Car Dashboard

Figure 4: SAM Car Dashboard

The Arrow Connect Cloud offers an API to enable third-party applications to visualize, integrate and analyze the data coming from the SAM Car. For instance, as shown in Figure 4, Arrow created a dynamic visual dashboard to display the data collection for spectators, pit crew, engineering teams and anyone interested in reviewing the technical performance of the car. The display features vehicle data collected from the main systems, including speed (top speed, average speed), steering degrees and acceleration. The dashboard also displays information from the subsystems, including in-vehicle temperature, humidity and light, along with elements of FMEA diagnostics data.


Figure 5: Sam Schmidt

Figure 5: Sam Schmidt

Remarkably, in May 2016, Sam achieved 152 mph in the SAM Car during practice laps at the Indianapolis 500. And in June 2016, he completed the Broadmoor Pikes Peak International Hill Climb in Colorado, famed for its 4,725 ft. elevation gain and 156 hairpin turns and twists over the 12.42 mile course, managing to reach 80 mph on some straightaways.

Summary: more than just a car

While the precision racecar represents a significant innovation and a very cool story the bigger story here is now the coming Internet of Things, not Internet of People, applications can reshape the fundamental infrastructure for the planet – transportation, energy, water, agriculture, construction, and healthcare.  If you’d like to see even more cases check out my new book, Precision: Principals, Practices and Solutions for the Internet of Things and the online class Precision: The Class.