Pages

Wednesday, December 5, 2018

Biped Gait Analysis

Human Gait Cycle 

To attempt to replicate human gait I first need to understand the stages of a normal bipedal gait cycle.

“A typical gait involves one foot placed forward with the second placed the same distance beyond the first” (Kong and Tomizuka, 2008, p. 3678).
This process repeated is the human gait.



“The repeating gait cycle can be subdivided into two periods (stance and swing)” (Handžić and Reed, 2015, p. 3). This is shown in figure 1.

Figure 1: Stance and swing periods of gait cycle (Kharb et al., 2011, p. 78)

“Stance is the term used to designate the entire period during which the foot is on the ground … swing applies to the time the foot is in the air for limb advancement” (Kharb et al., 2011, p. 78)

This is further divided into sub-stages. All sources describe the same stages during the gait cycle, though they differ in the level of subdivisions and the terminology used to reference each sub-stage.
An interpretation of the human gait cycle is shown in figure 2.

Figure 2: Stages of gait cycle (Kharb et al., 2011, p. 79)


Ensuring Natural Motion 

To replicate the sub-stages of the human gait cycle as closely as possible, there are a few more specific behaviours discussed in several papers which I need to remain aware of.

“In a healthy gait pattern, both legs move symmetrically and mirror all dynamics 180° out of phase" (Handžić and Reed, 2015, p. 3 cited Whittle, 2012).

Figure 3: Biped hip, knee and foot trajectory in sagittal view (Bajrami et al., 2016, p. 93)

My solution will need to contain logic which enforces the order of the gait cycle, maintaining the sub-stage ratios as the biped navigates varied terrain and obstacles.
Normal sub-stage proportions are shown in figure 4, where “the stance phase usually lasts about 60% of the cycle, the swing phase about 40% and each period of double support about 10%.” (Kharb et al., 2011, p. 79).

Figure 4: Phase durations (Kharb et al., 2011, p. 79)

The joint angles of the leg also need to be considered throughout the cycle. During the stance period the supporting leg will be straightened as the body moves over it, the limb in swing period will also straighten (knee lock) as it reaches toward the heel strike position. This behaviour is noted by Handžić and Reed (2015, p. 3), described in Figure 5.

Figure 5: Gait phases of a healthy walking human (Handžić and Reed, 2015, p. 3)

The joint angles of the upper body must also be considered during locomotion.
Research papers often mention the position of the spine in leaning into the direction of motion during a turn.
The relationship between the motion cycle of the upper limbs and their opposite leg counterparts are also often discussed: “there is more bending of elbows at larger speeds” (Shahid et al., 2012, p. 5). This will be another behaviour to manage as the locomotion speed of the biped alters.


Applying Gait Analysis To My Solution 

Understanding the sub-stages of the human gait cycle has provided a good foundation for the logic of my solution.

As stated by Handžić and Reed: “The upper body, which includes head, neck, trunk, and arms, moves along as a unit and is considered the passenger unit to the locomotor system, which consists of the legs” (2015, p. 3 cited Perry, 2010).
To mimic human locomotion, I will focus primarily on the interactions of the leg limbs with the terrain and navigating obstacles as they propel the hip joint forward along the sagittal plane. I will not attempt to simulate the model supporting itself against true gravity due to the scope of the project.
The spine and arms will be child chains of the hip joint with their behaviours being in response to those of the leg limbs.

I expect that a successful IK implementation combined which checks against a selection of gait sub-stages will be an effective way to produce the desired behaviour.
My first attempt will use Unity’s collision detection to advance through four stages:
  • Initial contact
  • Double support
  • Rear foot lift
  • Single support/swing
If this cycle proves to have insufficient granularity I will revisit my analysis to design an alternate approach.

As I will not be making my model respond directly to gravity and plan to effect motion based on limb end effector collision detection, I will limit the biped locomotion to a walking gait. A running gait involves periods without any ground contact as opposed to walking where “individuals retain ground contact throughout the gait cycle” (Handžić and Reed, 2015, p. 3 cited Whittle, 2012).

While I aim to include variation in walking velocity, I will not attempt to progress to a running gait unless walking behaviours have been implemented successfully and there is an appropriate amount of time remaining to reasonably create the new functionality.



References
Bajrami, X., Dermaku, A., Likaj, R., Demaku, N., Kikaj, A., Maloku, S. and Kikaj, D. (2016). Trajectory Planning and Inverse Kinematics Solver for Real Biped Robot with 10 DOF-s. IFAC-PapersOnLine, [online] 49(29), pp.88-93. Available at: https://www.sciencedirect.com/science/article/pii/S2405896316325460?via%3Dihub [Accessed 2 Dec. 2018].

Handžić, I. and Reed, K. (2015). Perception of gait patterns that deviate from normal and symmetric biped locomotion. Frontiers in Psychology, [online] 6. Available at: https://www.frontiersin.org/articles/10.3389/fpsyg.2015.00199/full [Accessed 4 Dec. 2018].

Kharb, A., Saini, V., Jain, Y. and Dhiman, S. (2011). A REVIEW OF GAIT CYCLE AND ITS PARAMETERS. [online] pp.78 - 83. Available at: https://www.researchgate.net/publication/268423123_A_review_of_gait_cycle_and_its_parameters [Accessed 1 Dec. 2018].

Kong, K. and Tomizuka, M. (2008). Smooth and Continuous Human Gait Phase Detection Based on Foot Pressure Patterns. 2008 IEEE International Conference onRobotics and Automation, [online] pp.3678 - 3689. Available at: https://ieeexplore-ieee-org.uos.idm.oclc.org/stamp/stamp.jsp?tp=&arnumber=4543775&tag=1 [Accessed 3 Dec. 2018].

Shahid, S., Nandy, N., Mondal, M., Ahamad, A., Chakraborty, C. and Nandi, G. (2012). A Study on Human Gait Analysis. Proceedings of the Second International Conference on Computational Science, Engineering and Information Technology, [online] pp.358 - 362. Available at: https://www.researchgate.net/publication/233792965_A_study_on_human_gait_analysis [Accessed 2 Dec. 2018].

Bibliography
Five Easy Pieces: Energetics of Human Locomotion. (2012). [video] Directed by dynamicwalking2012. Dynamic Walking Conference 2012 in Pensacola Beach, Florida: www.dynamicwalking.org. Available at: https://www.youtube.com/watch?v=mgyGS02_Bcg&list=PL3EBA143DC30E11C3&index=3

Gait Analysis and the Gait Cycle. (2017). [video] Directed by Morgan E. Available at: https://www.youtube.com/watch?v=5mDbF1zHHjw&t=8

Gait Cycle & Gait Analysis. (2015). [video] Directed by Physiotutors. Available at: https://www.youtube.com/watch?v=1u6d1CX7o9c&t=112s

Goswami, D. (2009). BIPED LOCOMOTION: STABILITY ANALYSIS, GAITGENERATION AND CONTROL. Doctor of Philosphy, Department of Electrical and Computer Engineering. NATIONAL UNIVERSITY OF SINGAPORE.

Perry, J. (2010). Gait Analysis: Normal and Pathological Function, 2nd Edn., Vol. 50.

Phases of Gait. (2008). [video] Directed by MedicalAnimaticsLLC. Available at: https://www.youtube.com/watch?v=_xq47GRr4Ss&list=PL3EBA143DC30E11C3&index=5

Whittle, M. (2012). Gait Analysis, 5th Edn. London: Elsevier Health Sciences.