Picture Gallery (2001-2003)

Most of the images below are "raw" and unrefined. For this gallery, we selected the pictures that gave the most accurate view of our work, rather than the prettiest.

Sensors

Here are some early Ladybug optic flow sensors.

This is our most recent Ladybug optic flow sensor. This sensor weights 4.5 grams, measures optic flow in excess of 30 radians per second, and responds to contrasts below two percent. This was the main "workhorse" sensor used in flight control research efforts in 2002 and 2003.

This is a newer (late 2003) version of the Ladybug sensor with improved optics

Aircraft Test bed

2003 Photos

This is one of our 100-gram aircraft

Three sensors mounted on the aircraft, ready for basic obstacle avoidance

2001-2002 Photos

This photo, circa June 2002, shows a complete system for testing biomimetic flight control algorithms. A laptop computer serves as the "base station" and is used to record sensor and control data that is down linked during flight. In the foreground are sensors, reflex boards, and other peripheral electronics. In the rear is a virgin Wingo!

This is a current "reflex board" that serves as the autopilot, control servo signal generator, and overall "glue" board.

This is our airplane hanger... It is often said that the greatest advantage of UAVs is that you can tell your squadron has arrived by the UPS truck pulling into the driveway!

Getting Ready for Flight

These photos will give you a sense of how we install the sensors and control electronics before a flight test. Typically we set up the aircraft the evening before a flight test, and then get up very early the next morning. Many of us are avid coffee drinkers...

Here you see a single sensor being installed in the Wingo fuselage in preparation for a flight experiment. The Wingo is made of foam, which allows us to cut or modify the structure for easy installation of various items.

Here is the downward optic flow sensor, ready to perform altitude control.

In one set of obstacle avoidance experiments, three sensors are placed along the yaw plane. The forward sensor measures yaw rate via optic flow. The left and right sensors are for detecting obstacles. This arrangement does not sample the entire visual field in the forward direction, yet it is able to detect larger obstacles on a collision course with the aircraft. The mount holding the sensors is not firmly mounted in the fuselage- it slips, which saves the sensors from mechanical shock during collisions.

This is the same sensor arrangement, with a view from the top.

Next we add the plastic silver "bra" to protect the aircraft's nose. We then add the rest of the electronics and do final testing. This includes looking at the serial bus to ensure that all items are properly communicating.

The next day, our UAV is taped up and ready to go! This setup looks ugly, but it works quite well.

Other Photographs

Here are a few more photographs.

When we were just experimenting with altitude control, we simply mounted a sensor on the wing. Inside the fuselage, we would have to place heavier items on the opposite side as the sensor to balance out the weight.

In our very first attempts at obstacle avoidance we mounted two sensors- one for each side. In this setup, we did not aim the aircraft at obstacles head-on. Instead, we just flew the aircraft towards obstacles at a shallow angle or from the side, and programmed the rudder to turn away when an obstacle was detected.

This is another obstacle avoidance setup from mid-2002. The board hanging off on the left is the "parameter board", essentially a "board-of-pots" used to set control algorithm parameters. The device on top is an Analog Devices MEMS gyro used for measuring yaw rate.

USAGE: All pictures may be copied and incorporated in other presentations, documents, videos, and web sites, provided that proper credit is given to Centeye, Inc.