Services

Simplify Robotics offers a variety of engineering services in mechanical, control, and software engineering.  Our core competency is in robotics.  In particular, we are experienced in manipulator kinematics, dynamics, control, and simulations.  We can also lead or participate in mechanical design projects.  Our in-house CAD system of choice is SolidWorks.  Furthermore, we can lead or participate in software development projects.  Our programming expertise is with C/Matlab/LabVIEW, and our platform of choice is Linux/Unix.  Below is a partial list of our past projects, but it does NOT comprehensively represent our capabilities.  For more information, please visit our contact page.

Past Projects

The Human Spaceflight programs from around the globe use a number of manipulators designed to assist astronauts in performing their tasks in orbit.  These include the 6-jointed Shuttle Remote Manipulator System (SRMS), the 7-jointed Space Station Remote Manipulator System (SSRMS), the Special Purpose Dexterous Manipulator (SPDM) with two 7-jointed arms, the 7-jointed European Robotic Arm (ERA), and the 7-jointed Japanese Experimental Module Remote Manipulator System (JEMRMS).  Simplify Robotics developed a lightweight manipulator kinematics calculation utility for the Automation, Robotics, and Simulation Division (AR&SD) at NASA Johnson Space Center, for the purpose of providing a uniform and independent capability for testing the kinematics implementations of these robots, in the multitude of simulations for analysis, training, and planning.

Simplify Robotics assisted in troubleshooting a newly developed simulation of the JEMRMS manipulator at AR&SD in NASA.  This involved testing the operation of the individual features of the control interface, and then helping to debug any problems that arose.  The simulations were written in C, developed using the Trick simulation development environment, and ran on the Linux Operating System.  The control interface has dozens of individual windows, each is similar to the one shown to the left.

Following the Space Shuttle Columbia disaster in 2003, NASA launched several investigations and devised many contingency scenarios, during their return to flight activities.  One of them involved an Orbiter Repair Maneuver (ORM), which was designed to position an astronaut underneath the Shuttle's Orbiter, where (s)he would have access to any damaged heat shielding tiles.  The maneuver entailed the Orbiter grappling the International Space Station (ISS) using the Shuttle's arm, while positioning the astronaut at the tip of the Station's arm underneath the Orbiter.  This maneuver required connecting large masses (Orbiter and Station) by thin, slender, flexible manipulator links.  Simplify Robotics performed an analysis of this dynamic system.  This involved the use of Autolev analysis software.

The Robonaut is a humanoid robot meant to be an assistant for an astronaut in orbit.  Its humanoid form factor enables it to fit in many of the same settings in which an astronaut would as well.  Its humanoid hands enable it to mechanically handle the same tools that would be used by an astronaut.  There are many researchers in various universities and organizations who contribute software and algorithms to the project.  In order to facilitate their work, the Robosim, a Windows based simulation has been developed, which can be distributed among all project contributors.  Additionally, there is an API which enables an external program to communicate with either the physical Robonaut or with Robosim.  Simplify Robotics ported and tested the Robonaut API from the Windows to the Linux platform.

A number of tactile sensors have been developed at AR&SD in NASA.  In order to adequately test various features on these sensors, Simplify Robotics designed a fixture to constrain the target sensors, and to apply variable contact loads.  The resulting design also provided capabilities for fine positioning of the target, directional control of contact loads, and fine control of contact load magnitudes.

Testimonial

For a 3-dimensional End-Effector (EE) which results in m=6 kinematic equations (described below under Inverse Kinematics), driven by an n=7 revolute jointed arm, there are more variables than equations.  This is an underdetermined case which typically results in an infinite number of solutions.  The 7-jointed manipulator can traverse through the continuum of solution joint configurations, while the position and the orientation of the EE remain fixed.  This is called self-motion.  Simplify Robotics developed a software utility for analysis of this self-motion for the SSRMS and the SPDM manipulators.

One motivation for the use of robots is for manipulation of objects.  A multifingered hand is versatile in grasping objects, when their shapes are unpredictable and without adequate features for mating with more traditional end-effectors.  However, multifingered grasping can be a challenge.  One such challenge is to constrain a payload such that external forces and torques (within a defined magnitude) do not disturb the grasped payload.  This is a wrench closure condition.  Simplify Robotics investigated this problem, and similar peripheral ones.  The result was a mathematical framework for addressing all grasp problems, including the wrench closure condition, optimal grasping, multipayload grasping, and gaited grasping.

Testimonial Publication

In 3-dimensional space, there are 3 dimensions for achievable positions (X,Y,Z), and 3 for orientations (e.g.: about X,Y,Z), for a total of m=6 degrees of manipulability.  In order for a manipulator to be capable of positioning and orienting its End-Effector (EE) arbitrarily, it is necessary for it to have at least n=6 Degrees-of-Freedom (DOF).  In many applications, it is necessary to determine the joint values which result in the correct positioning and orienting of the EE.  This analysis is called Inverse Kinematics.  Formally, it can be presented as m=6 kinematic equations, in terms of n joint variables.  For typical 6-jointed manipulators which employ rotary or revolute joints, these 6 equations have 6 unknowns, and are highly nonlinear, coupled, and complex.  Often, there are multiple distinct solutions.  Typical inverse kinematics algorithms are numerical in nature, starting with an initial guess, and converging toward only one of the possible solutions.

NASA and other international space organizations employ several n=7 revolute jointed manipulators.  Simplify Robotics investigated a technique which would produce all solutions for a 6 revolute jointed arm, or its equivalent, a 7 revolute jointed arm with one joint arbitrarily held fixed.  This scenario arises when one of the joints of a 7 jointed space manipulator becomes damaged and no longer functions.  This investigation involved the use of symbolic algebra computations, and the GiNaC C++ library for Computer Algebra.

Under water at NASA Johnson Space Center's Neutral Buoyancy Lab (NBL), astronauts are trained for Extra-Vehicular Activity (EVA).  Many training operations involve underwater mock-ups of the SRMS and the SSRMS.  Occasionally, one underwater manipulator may be out of service.  To avoid loss of training time, Simplify Robotics developed a control software plug-in that can be used to drive the 6-DOF SRMS underwater manipulator with the control workstation of the 7-DOF SSRMS.  In this scenario, the operator drives the existing simulation of the 7-DOF SSRMS which acts as the master, and the 6-DOF SRMS acting as the slave, moves to match the position and orientation of the master's EE with its own.  The bottom figure on the left shows the two robots with the master-slave coupling.  The bases of the two robots may be at different positions and orientations.

Testimonial(1) , Testimonial(2)

The Canadian built Special Purpose Dexterous Manipulator (SPDM) is one of many space manipulators available to NASA.  For analysis, planning, and training, simulations of this manipulator are necessary.  The end-effector of each SPDM arm is the ORU Tool Changeout Mechanism (OTCM).  Simplify Robotics developed the simulation math models which kinematically mimic the behavior of the OTCMs.  The math models were developed in C using the Trick simulation development environment.

This was a sequel to the earlier successful project (see below), in which the dynamics of the cable and spring mechanisms inside the end-effector (EE) of the Space Station's manipulator were modeled.  In this case, a higher fidelity model was pursued, one which required the development and use of a numerical optimization library.  The math models were developed in C using the Trick simulation development environment.

In order to perform assembly missions in orbit, astronauts must be trained in manipulator operations.  The trainees can benefit by first learning about generic manipulator operations prior to specialization with a particular manipulator.  Therefore, starting with a partially developed simulation from NASA, Simplify Robotics joined efforts with Titan-LinCom (now L-3) to fully develop and deliver the training simulation as specified by NASA requirements.  The simulation was developed using C in the Trick simulation development environment.  Our contributions included developing closed form 6-DOF inverse kinematics, self-collision detection, and enhancing existing motion control and other portions of the software.  The math models were developed in C using the Trick simulation development environment.

The SSRMS grapples payloads using rotary snare and linear carriage mechanisms inside its Latching End Effector (LEE). For design, analysis, and crew training purposes, Simplify Robotics developed a simulation math model which characterizes the dynamics of the LEE's complex cable and spring system and its effects on a space payload with a Grapple Fixture (GF). The math models were implemented using C in the Trick simulation development environment, which is heavily used at NASA Johnson Space Center. The inertial, gearing, and motor dynamics of the rotary snare and linear carriage were also prototyped using Matlab.

Testimonial

Motion planning for a manipulator is more difficult than motion planning for a mobile robot:  (1) Mobile agents constrained to the surface of a 3-dimensional terrain are typically operating with 2-dimensions of control (steering and throttle/brake) and in a 3-dimensional workspace (2D for position, 1D for orientation).  In contrast, typical manipulators have 6-7 joints, hence they operate in a 6-7 dimensional joint space, and their workspace is 6-dimensional (3D position, 3D orientation).  Since computation complexity grows exponentially with the dimension of the search space, the search for a path for a manipulator can be orders of magnitude more costly than the same for a mobile agent.  (2) Since manipulator joints are typically revolute and the kinematics involve highly coupled trigonometric equations, the mapping between a manipulator's joint- and work- spaces is much more complex than the equivalent for mobile agents.  (3) For mobile agents, there is typically only a single body to consider.  This single body is to avoid obstacles, given its motion constraints.  In contrast, planning for a manipulator requires motion considerations for several link bodies, while each avoids collisions with obstacles, and while each adheres to physical constraints associated with its consecutive joints and links.

During his Ph.D. research, the founder, Arjang Hourtash, spent several years studying automated motion planning for manipulators.  The result was a divide and conquer approach which dimensionally decomposes the joint space, hence reducing the computation cost of the search for a solution path.  Several aspects of this approach were investigated thoroughly, including the optimization of the automation parameters.

Testimonials

Contact