Posts Tagged ‘William (Red) Whittaker’

1992-4 – Dante & Dante II – John E. Bares & William “Red” Whittaker (American)


In 1992 a walking robot named Dante1 was designed and built at Carnegie Mellon University. Using a tensioned tether, Dante can ascend and descend steep slopes. It is designed to rappel into and explore active volcanic craters. The Dante project was an ambitious attempt to proceed, in 10 months, from idea to implementation. The culmination was an expedition to an active volcano, Mount Erebus, in Antarctica.
In terms of robot science, the objectives were to demonstrate a real exploration mission, rough terrain locomotion, environmental survival, and selfsustained operation in the cold, windy, bright, rugged Antarctic environment. The expedition to Mount Erebus was indeed a real mission. Dante climbed into the steep sided crater, surviving the environment and operating in a self-sustaining manner until the failure of a critical component, the communications tether, necessitated rescue and the premature termination of the crater exploration. The volcano science objectives were also real and involved the study of Mount Erebus’ unique convecting magma lake. Dante was equipped with instruments to determine the chemical and isotopic composition of gas generated by the volcano’s magma lake, to measure the radioactivity of materials near the lake and to measure the temperature of the magma itself.
Dante can rappel up and down steep slopes and surmount obstacles as large as 1 meter in height. It can perceive and model the terrain around it using a scanning laser rangefinder and a trinocular camera system. Its planning software determines safe paths and adjusts the gait to avoid obstacles. While some of its computers reside off-board at a base station and are linked via a fiber optic cable, Dante is computationally self-sufficient, able to chart its own course, react to perceived terrain, and acquire data from science payload sensors. Dante’s eight pantographic legs are organized in frames of four—an inner frame and an outer frame—with each coupled through a drivetrain that provides an intrinsic walking motion. To walk, four legs simultaneously lift and reach forward while the other four supporting legs propel the body. Each of the legs can individually adjust its height to avoid obstacles in the terrain. On steep slopes the tensioned tether provides a reactive force to gravity, assists in maintaining equilibrium, and allows Dante to rappel like a mountain climber.

1. The robot is named Dante in reference to the poem, The Divine Comedy by Dante Alegheri, in which Dante travels to the underworld. Erebus is a cloud of mist that obscures the entry to hell. Early mission scenarios also had a transport robot named Virgil for the Roman poet who guides Dante during his quest. In keeping with the theme, the cart used to carry Dante around is called Geryon after a flying daemon who gave Dante a lift.

See Dante pdf here.

See Electronics Australia Jan 1994 article here .

Dante II:

The CMU Field Robotics Center (FRC) developed Dante II, a tethered walking robot, which explored the Mt. Spurr (Aleutian Range, Alaska) volcano in July 1994. High-temperature, fumarole gas samples are prized by volcanic science, yet their sampling poses significant challenge. In 1993, eight volcanologists were killed in two separate events while sampling and monitoring volcanoes. The use of robotic explorers, such as Dante II, opens a new era in field techniques by enabling scientists to remotely conduct research and exploration.
Using its tether cable anchored at the crater rim, Dante II is able to descend down sheer crater walls in a rappelling-like manner to gather and analyze high temperature gasses from the crater floor. In addition to contributing to volcanic science, a primary objective of the Dante II program is to demonstrate robotic exploration of extreme (i.e., harsh, barren, steep) terrains such as those found on planetary surfaces.

Dante II Technical Information
•Type: Eight-legged rappelling frame walker
•Manufacturer: Field Robotics Center, Carnegie Mellon University
•List price: $1,700,000 (destination charges not included)
General Data
•Length: 120 in. (up/downhill)
•Width: 85 in. (cross-slope)
•Height: 120 in. (includes sensor mast)
•Tether: 1000 ft. (length), 0.45 in. (diameter)
•Weight: 1700 lb.
•Leg type: 4x pantograph
•Speed: 1 m/min.
•Stride: 45 in. (single-step)
•Max. turn: 11 deg. (single-step)
•Max. obstacle clearing: 50 in.
•Max. rapelling distance: 1000 ft.
•Temp. range: 0 to 100 degrees F
•Leg: position (0.1 in. resolution), axial force (0 to 3000 lb at pantograph mount)
•Tether: tension (0 to 3000 lb.), exit angle (-90 to 90 degree, two-axis), payout (30 cm resolution)
•Body: Two-axis inclinometers (0.1 degree, pitch & roll)
•Single gas sensors: H2S, SO2, CO2
•Temperature: thermocouple (foot mounted), ambient
•Cameras: Color stereo pair (pan/tilt mounted), Color zoom (pan/tilt mounted), Kodak DCS-100 digital 35mm camera (pan/tilt mounted), Four leg view cameras (bottom frame mounted)
•Scanning laser rangerfinder
Computing & Electronics
•Computing: VME single board computers (Sparc, M68030's), VxWorks operating system
•Power: 3A at 1000VAC (peak), 2 KW (nominal)
•Two VTEL video codecs
•1024 Kbit/sec satellite link (includes 192 Kbit/sec TCP/IP, two 384 Kbit/sec digital video) 01/08/2012 00:51:22

See list of Dante II pdf's here.

See 1min 45sec into video clip to see Dante II. Thanks to David Buckley for locating this for me.

1988-91 – AMBLER – John Bares & William “Red” Whittaker (American)

A good example of the "big iron" approach to mobile robots is AMBLER (acronym for Autonomous MoBiLe Exploration Robot), developed by Carnegie Mellon University and the Jet Propulsion Laboratory. This behemoth stands about 5m (16.4ft) tall, is up to 7m (23.0ft) wide, and weights 2500 kg (5512 lb). It moves at a blistering 35 cm (13.8 in) per minute. Just sitting still, it consumes 1400 W of power. Ask it to walk and it sucks up just about 4000 W.

AMBLER showing time-lapse traces on one leg.

Some of the Carnegie-Mellon team with AMBLER highlight its immense size.

See a few pdf's describing AMBLER mainly here and here.

Below text sourced from here.

Ambler: Autonomous Orthogonal Legged Walking Robot

The Ambler robot was designed for walking under the particular constraints of planetary terrain, where there are meter-sized boulders, deep crevices, and steep slopes-an altogether inhospitable environment that defies humans and wheeled machines alike. Therefore, the six-legged Ambler travels over extremely rugged terrain without the close aid of humans. Autonomously, the Ambler builds detailed terrain maps; plans its own sequence and location of steps; and controls its movement, balance, and stability. In extensive tests, the Ambler has traveled thousands of meters, taken thousands of steps, and negotiated terrains that defy other robots.


Ambler walks like no other machine and like no other creature in nature: Stepping with any leg in any sequence, the Ambler has the patented capability to move its rear-most leg past all other legs in order to travel over extreme terrain as efficiently as possible. Also, while most animals bend their legs to step and walk, Ambler's legs remain vertical, while they swing horizontally, then lengthen themselves vertically, like a telescope, to touch the ground. Such legs do not rock or sway in the act of stepping, thus risking unnecessary collision with obstacles. More flexible, animal-like legs require substantially more sensing and planning from a robot, but the Ambler's unbendable legs decrease both the consequences and the extra planning that would be necessary for bendable legs.

The robot's height of 3.5 meters enables it to step over obstacles as high as one meter. At the same time, no matter how rough the terrain, the Ambler walks upright, keeping its legs vertical and its body horizontalÑand keeping its laser rangefinder steady. It is through data from the laser rangefinder that the Ambler's perception system builds computerized maps of the terrain. (See Terrain Mapping, Krotkov.) In fact, Ambler's walking design facilitates perception of the terrain by maintaining a steady and level posture (on a 30 degree slope). When the robotÕs perception system merges laser images from different viewpoints into a larger composite picture of the terrain, the robot's stability gives its laser images a good registrationÑthat is, leaves very little unintended overlap and no gaps between the various image viewpoints. The robot's height also gives the laser rangefinder a high-vantage with which to better view the terrain, and promotes a high quality of sensor data.


Although remote human operators tell the Ambler where to go, the robot itself plans the steps it must take to get there (see also, Gait Configuration of Legged Robots, Wettergreen). The robot's gait planner takes into account not only terrain constraints but also its own walking capabilities: how far the robot's legs can reach, how long the legs can extend, how far the robot's body can stray from its center of gravity, where the robot can move each leg without colliding into another leg, and how it can place its legs so that its body-which moves alternately with the legs-also has a clear path to move forward.

After the gait planner has intersected all of these constraints and determined a limited number of steps, the footfall planner considers which of the available steps offer the best footholds and are more efficient in time and energy. The footfall planner has learned, through a neural network, which footholds are optimal, having been presented examples during its development of good and bad footholds. The leg recovery planner finally determines how to move each leg without colliding into something mid-move. At the same time, the robot's planners must weigh the various constraints. For example, the robot's body must move as far forward as possible (to increase efficiency of speed), without moving beyond its center of gravity.


Getting the various on-board systems to interact efficiently involves the use of Task Control architecture (see Task-Level Communication and Control, Simmons). Like a switch-board operator, TCA facilitates communication between the Ambler's various systems, coordinates the robot's plans, sequences tasks, and monitors actions and recovers from problems. Task Control Architecture enables planning, perception, and real-time control to work concurrently.

Early impression by the artist Erik Viktor.

Tags: , , , , , , ,