Is repeated measures ANOVA appropriate for my data set?

My project involves designing various controllers of a biomechanical system, and then once the controller design is complete, testing each controller on a single, standard set of 500 tasks, and determining the performance of each controller on this single set of tasks, so that I can compare the controllers' performance against each other. The performance metrics I use are all quantitative, and include position error in centimeters & orientation error in degrees. I have a total of 5 controllers that I'm analyzing.

Up until now, my analysis has consisted of the following:
~ checked for normal data distribution using the Kolmogorov-Smirnov test --> the data failed
~ checked for homogeneity of variance using Levene's test --> the data failed
~ concluded that due to the above 2 results, performing a 1-way ANOVA would not be possible
~ applied nonparametric Kruskal-Wallis ANOVA analysis to compare the 5 controllers' performance on the single 500-task set

I recently spoke with a project collaborator who suggested that, since all 5 controllers are being tested on the same 500-task set, then I should instead be using a repeated measures ANOVA analysis instead of Kruskal-Wallis ANOVA.

However, from the reading I've done about repeated measures ANOVA, it seems that this test should be applied only when the SAME subject/controller performs the same task MORE than once, which is not true for my project; again, each controller is applied to the same 500-task set only *once*.

Before I reply to my collaborator, I'm hoping to get a confirmation that, indeed, repeated measures ANOVA isn't appropriate to use for my project, and that Kruskal-Wallis ANOVA is the correct method to use here. Thanks in advance for your guidance!