
Control and Evaluation of a Humanoid Robot with Rolling Contact Knees
Seung Hyeon Bang1, Carlos Gonzalez1, Junhyeok Ahn2, Nicholas Paine3and Luis Sentis1
Abstract— In this paper, we introduce the humanoid robot
DRACO 3 by providing a high-level description of its design and
control. This robot features proximal actuation and mechanical
artifacts to provide a high range of hip, knee and ankle motion.
Its versatile design brings interesting problems as it requires
a more elaborate control system to perform its motions. For
this reason, we introduce a whole body controller (WBC) with
support for rolling contact joints and show how it can be easily
integrated into our previously presented open-source Planning
and Control (PnC) framework. We then validate our controller
experimentally on DRACO 3 by showing preliminary results
carrying out two postural tasks. Lastly, we analyze the impact
of the proximal actuation design and show where it stands in
comparison to other adult-size humanoids.
I. INTRODUCTION
Dynamic motions for legged systems require accurate
sensing of the robot’s states and high performance con-
trol. To realize the latter, several important issues must be
considered. On one hand, efficient transmissions with low
friction/stiction and low backlash are desired. In addition,
mechanical designs capable of achieving a wide range of
motion (RoM) are important and non-trivial to achieve. On
the other hand, these types of mechanisms come at the ex-
pense of an increase in complexity of the controllers needed
to exploit the potential of the high-dimensional system.
Humanoids have often employed collocated actuation (mo-
tors directly located at each joint) because of its simplicity
in design [1], [2]. However, the performance of these robots
degrades when using simplified models for planning because
of model discrepancy caused by the heavy distal mass on
their legs [3]. Due to this, their design has shifted in favor
of proximal actuation (placing heavy motors near the body)
to reduce the limbs’ distal mass for dynamic maneuvers. As
part of the solution to achieve this, a transmission system
that has been explored frequently in legged robots consists of
cable-based drive systems [4]–[6]. This is due to their light
weight as well as effective power transmission capability,
thus helping provide high torque density when incorporated
into the design. Additionally, this kind of transmission is
more efficient because the mechanical losses due to friction
are small and, thus, provides higher torque transparency.
This work was supported by ONR Grant# N000141512507.
1S.H. Bang, C. Gonzalez, and L. Sentis are with the Department of
Aerospace Engineering and Engineering Mechanics, The University of
Texas at Austin, TX 78712, USA bangsh0718@utexas.edu,
carlos.gonzalez@utexas.edu,
lsentis@austin.utexas.edu
2J. Ahn is with the Department of Mechanical Engineer-
ing, The University of Texas at Austin, TX, 78712, USA
junhyeokahn91@gmail.com
3N. Paine is with Apptronik, Inc., 110701 Stonehollow Dr STE 150,
Austin, TX, 78758, USA n.a.paine@gmail.com
1 DOF Neck
3 DOF Shoulder
1 DOF Elbow
2 DOF Wrist
3 DOF Hip
1 DOF Knee
2 DOF Ankle
(a) (b)
Fig. 1. DRACO 3 Humanoid. (a) Wireframe of DRACO 3 showing the
degrees of freedom on the right side of the robot, and their respective axis
of rotation. (b) DRACO 3 standing at full height.
In a similar manner, DRACO 3 (shown in Fig. 1) was
designed bearing a cable-based drive system in mind for
proximal actuation. In addition, a rolling contact joint (RCJ)
mechanism is employed to enhance its RoM. Although
RCJs have been widely used in other fields such as in
lower extremity exoskeleton design to reduce misalignment
between the human’s and the device’s joint [7], [8], and in
robotic fingers to realize large RoM as well as to decrease
internal friction [9], it has not been adopted on humanoids
due to its mechanical complexity and complicated control
design. By means of a RCJ on the knee, DRACO 3 achieved
proximal actuation and high ranges of motion.
Typical methods to control complex systems such as
legged robots often make use of a WBC, which compute
joint-level commands to achieve desired operational space
tasks. This type of controller is often designed by considering
a model of the robot along with multiple physical and
environmental constraints (e.g., friction). In addition, any
complexities in the joint mechanisms must be taken into
account either within or outside of the WBC module. The
recent work by [10] presents a rich literature review on
different types of WBCs. In this work, we chose to pursue
a WBC, similar to that in [10], [11] for its high flexibility
and ease of intuitive tuning. Previous works [7]–[9], [12]
have studied the control of complex joint mechanisms (e.g.,