Is Harvard or Carnegie Mellon better for undergraduate robotics?

For an undergraduate focused specifically on robot perception and autonomous systems, Carnegie Mellon is the more direct and deeply specialized choice. Its strengths in computer vision, machine learning, sensing, planning, manipulation, and field robotics align closely with the areas you named.

CMU suits a student who wants robotics to be the center of their academic life from the beginning. You can build a curriculum around robotics while drawing on strong computer science, mechanical engineering, electrical and computer engineering, and AI offerings. The large robotics community also makes it easier to find project teams, research groups, and advanced technical electives connected to autonomy and perception. The tradeoff is an academically intense, technology-focused environment where many classmates will be similarly specialized.

Harvard is compelling for a student who wants to study robotics within a broader liberal-arts and engineering setting. Through the Harvard John A. Paulson School of Engineering and Applied Sciences, students can pursue computer science or engineering sciences and connect robotics to AI, bioengineering, design, or policy. Harvard offers meaningful research opportunities, including work related to intelligent systems and manipulation, but it does not provide the same undergraduate robotics-specific infrastructure or breadth of dedicated coursework as CMU.

For your stated interests, CMU offers the stronger overall robotics education and research environment. Harvard becomes the more attractive option if you value a broader undergraduate experience and expect your interests to extend substantially beyond technical robotics.